?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'jc/diff-irreversible-delete' into next
[git/dscho.git] / diff.c
blobf795496b90a8cb447c130cff833b7d69f92827b2
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 git_config_rename(const char *var, const char *value)
70 {
71 if (!value)
72 return DIFF_DETECT_RENAME;
73 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
74 return DIFF_DETECT_COPY;
75 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
76 }
77
78 /*
79 * These are to give UI layer defaults.
80 * The core-level commands such as git-diff-files should
81 * never be affected by the setting of diff.renames
82 * the user happens to have in the configuration file.
83 */
84 int git_diff_ui_config(const char *var, const char *value, void *cb)
85 {
86 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
87 diff_use_color_default = git_config_colorbool(var, value, -1);
88 return 0;
89 }
90 if (!strcmp(var, "diff.renames")) {
91 diff_detect_rename_default = git_config_rename(var, value);
92 return 0;
93 }
94 if (!strcmp(var, "diff.autorefreshindex")) {
95 diff_auto_refresh_index = git_config_bool(var, value);
96 return 0;
97 }
98 if (!strcmp(var, "diff.mnemonicprefix")) {
99 diff_mnemonic_prefix = git_config_bool(var, value);
100 return 0;
101 }
102 if (!strcmp(var, "diff.noprefix")) {
103 diff_no_prefix = git_config_bool(var, value);
104 return 0;
105 }
106 if (!strcmp(var, "diff.external"))
107 return git_config_string(&external_diff_cmd_cfg, var, value);
108 if (!strcmp(var, "diff.wordregex"))
109 return git_config_string(&diff_word_regex_cfg, var, value);
110
111 if (!strcmp(var, "diff.ignoresubmodules"))
112 handle_ignore_submodules_arg(&default_diff_options, value);
113
114 return git_diff_basic_config(var, value, cb);
115 }
116
117 int git_diff_basic_config(const char *var, const char *value, void *cb)
118 {
119 if (!strcmp(var, "diff.renamelimit")) {
120 diff_rename_limit_default = git_config_int(var, value);
121 return 0;
122 }
123
124 switch (userdiff_config(var, value)) {
125 case 0: break;
126 case -1: return -1;
127 default: return 0;
128 }
129
130 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
131 int slot = parse_diff_color_slot(var, 11);
132 if (slot < 0)
133 return 0;
134 if (!value)
135 return config_error_nonbool(var);
136 color_parse(value, var, diff_colors[slot]);
137 return 0;
138 }
139
140 /* like GNU diff's --suppress-blank-empty option */
141 if (!strcmp(var, "diff.suppressblankempty") ||
142 /* for backwards compatibility */
143 !strcmp(var, "diff.suppress-blank-empty")) {
144 diff_suppress_blank_empty = git_config_bool(var, value);
145 return 0;
146 }
147
148 if (!prefixcmp(var, "submodule."))
149 return parse_submodule_config_option(var, value);
150
151 return git_color_default_config(var, value, cb);
152 }
153
154 static char *quote_two(const char *one, const char *two)
155 {
156 int need_one = quote_c_style(one, NULL, NULL, 1);
157 int need_two = quote_c_style(two, NULL, NULL, 1);
158 struct strbuf res = STRBUF_INIT;
159
160 if (need_one + need_two) {
161 strbuf_addch(&res, '"');
162 quote_c_style(one, &res, NULL, 1);
163 quote_c_style(two, &res, NULL, 1);
164 strbuf_addch(&res, '"');
165 } else {
166 strbuf_addstr(&res, one);
167 strbuf_addstr(&res, two);
168 }
169 return strbuf_detach(&res, NULL);
170 }
171
172 static const char *external_diff(void)
173 {
174 static const char *external_diff_cmd = NULL;
175 static int done_preparing = 0;
176
177 if (done_preparing)
178 return external_diff_cmd;
179 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
180 if (!external_diff_cmd)
181 external_diff_cmd = external_diff_cmd_cfg;
182 done_preparing = 1;
183 return external_diff_cmd;
184 }
185
186 static struct diff_tempfile {
187 const char *name; /* filename external diff should read from */
188 char hex[41];
189 char mode[10];
190 char tmp_path[PATH_MAX];
191 } diff_temp[2];
192
193 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
194
195 struct emit_callback {
196 int color_diff;
197 unsigned ws_rule;
198 int blank_at_eof_in_preimage;
199 int blank_at_eof_in_postimage;
200 int lno_in_preimage;
201 int lno_in_postimage;
202 sane_truncate_fn truncate;
203 const char **label_path;
204 struct diff_words_data *diff_words;
205 struct diff_options *opt;
206 int *found_changesp;
207 struct strbuf *header;
208 };
209
210 static int count_lines(const char *data, int size)
211 {
212 int count, ch, completely_empty = 1, nl_just_seen = 0;
213 count = 0;
214 while (0 < size--) {
215 ch = *data++;
216 if (ch == '\n') {
217 count++;
218 nl_just_seen = 1;
219 completely_empty = 0;
220 }
221 else {
222 nl_just_seen = 0;
223 completely_empty = 0;
224 }
225 }
226 if (completely_empty)
227 return 0;
228 if (!nl_just_seen)
229 count++; /* no trailing newline */
230 return count;
231 }
232
233 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
234 {
235 if (!DIFF_FILE_VALID(one)) {
236 mf->ptr = (char *)""; /* does not matter */
237 mf->size = 0;
238 return 0;
239 }
240 else if (diff_populate_filespec(one, 0))
241 return -1;
242
243 mf->ptr = one->data;
244 mf->size = one->size;
245 return 0;
246 }
247
248 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
249 static unsigned long diff_filespec_size(struct diff_filespec *one)
250 {
251 if (!DIFF_FILE_VALID(one))
252 return 0;
253 diff_populate_filespec(one, 1);
254 return one->size;
255 }
256
257 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
258 {
259 char *ptr = mf->ptr;
260 long size = mf->size;
261 int cnt = 0;
262
263 if (!size)
264 return cnt;
265 ptr += size - 1; /* pointing at the very end */
266 if (*ptr != '\n')
267 ; /* incomplete line */
268 else
269 ptr--; /* skip the last LF */
270 while (mf->ptr < ptr) {
271 char *prev_eol;
272 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
273 if (*prev_eol == '\n')
274 break;
275 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
276 break;
277 cnt++;
278 ptr = prev_eol - 1;
279 }
280 return cnt;
281 }
282
283 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
284 struct emit_callback *ecbdata)
285 {
286 int l1, l2, at;
287 unsigned ws_rule = ecbdata->ws_rule;
288 l1 = count_trailing_blank(mf1, ws_rule);
289 l2 = count_trailing_blank(mf2, ws_rule);
290 if (l2 <= l1) {
291 ecbdata->blank_at_eof_in_preimage = 0;
292 ecbdata->blank_at_eof_in_postimage = 0;
293 return;
294 }
295 at = count_lines(mf1->ptr, mf1->size);
296 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
297
298 at = count_lines(mf2->ptr, mf2->size);
299 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
300 }
301
302 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
303 int first, const char *line, int len)
304 {
305 int has_trailing_newline, has_trailing_carriage_return;
306 int nofirst;
307 FILE *file = o->file;
308
309 if (o->output_prefix) {
310 struct strbuf *msg = NULL;
311 msg = o->output_prefix(o, o->output_prefix_data);
312 assert(msg);
313 fwrite(msg->buf, msg->len, 1, file);
314 }
315
316 if (len == 0) {
317 has_trailing_newline = (first == '\n');
318 has_trailing_carriage_return = (!has_trailing_newline &&
319 (first == '\r'));
320 nofirst = has_trailing_newline || has_trailing_carriage_return;
321 } else {
322 has_trailing_newline = (len > 0 && line[len-1] == '\n');
323 if (has_trailing_newline)
324 len--;
325 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
326 if (has_trailing_carriage_return)
327 len--;
328 nofirst = 0;
329 }
330
331 if (len || !nofirst) {
332 fputs(set, file);
333 if (!nofirst)
334 fputc(first, file);
335 fwrite(line, len, 1, file);
336 fputs(reset, file);
337 }
338 if (has_trailing_carriage_return)
339 fputc('\r', file);
340 if (has_trailing_newline)
341 fputc('\n', file);
342 }
343
344 static void emit_line(struct diff_options *o, const char *set, const char *reset,
345 const char *line, int len)
346 {
347 emit_line_0(o, set, reset, line[0], line+1, len-1);
348 }
349
350 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
351 {
352 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
353 ecbdata->blank_at_eof_in_preimage &&
354 ecbdata->blank_at_eof_in_postimage &&
355 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
356 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
357 return 0;
358 return ws_blank_line(line, len, ecbdata->ws_rule);
359 }
360
361 static void emit_add_line(const char *reset,
362 struct emit_callback *ecbdata,
363 const char *line, int len)
364 {
365 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
366 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
367
368 if (!*ws)
369 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
370 else if (new_blank_line_at_eof(ecbdata, line, len))
371 /* Blank line at EOF - paint '+' as well */
372 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
373 else {
374 /* Emit just the prefix, then the rest. */
375 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
376 ws_check_emit(line, len, ecbdata->ws_rule,
377 ecbdata->opt->file, set, reset, ws);
378 }
379 }
380
381 static void emit_hunk_header(struct emit_callback *ecbdata,
382 const char *line, int len)
383 {
384 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
385 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
386 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
387 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
388 static const char atat[2] = { '@', '@' };
389 const char *cp, *ep;
390 struct strbuf msgbuf = STRBUF_INIT;
391 int org_len = len;
392 int i = 1;
393
394 /*
395 * As a hunk header must begin with "@@ -<old>, +<new> @@",
396 * it always is at least 10 bytes long.
397 */
398 if (len < 10 ||
399 memcmp(line, atat, 2) ||
400 !(ep = memmem(line + 2, len - 2, atat, 2))) {
401 emit_line(ecbdata->opt, plain, reset, line, len);
402 return;
403 }
404 ep += 2; /* skip over @@ */
405
406 /* The hunk header in fraginfo color */
407 strbuf_add(&msgbuf, frag, strlen(frag));
408 strbuf_add(&msgbuf, line, ep - line);
409 strbuf_add(&msgbuf, reset, strlen(reset));
410
411 /*
412 * trailing "\r\n"
413 */
414 for ( ; i < 3; i++)
415 if (line[len - i] == '\r' || line[len - i] == '\n')
416 len--;
417
418 /* blank before the func header */
419 for (cp = ep; ep - line < len; ep++)
420 if (*ep != ' ' && *ep != '\t')
421 break;
422 if (ep != cp) {
423 strbuf_add(&msgbuf, plain, strlen(plain));
424 strbuf_add(&msgbuf, cp, ep - cp);
425 strbuf_add(&msgbuf, reset, strlen(reset));
426 }
427
428 if (ep < line + len) {
429 strbuf_add(&msgbuf, func, strlen(func));
430 strbuf_add(&msgbuf, ep, line + len - ep);
431 strbuf_add(&msgbuf, reset, strlen(reset));
432 }
433
434 strbuf_add(&msgbuf, line + len, org_len - len);
435 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
436 strbuf_release(&msgbuf);
437 }
438
439 static struct diff_tempfile *claim_diff_tempfile(void) {
440 int i;
441 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
442 if (!diff_temp[i].name)
443 return diff_temp + i;
444 die("BUG: diff is failing to clean up its tempfiles");
445 }
446
447 static int remove_tempfile_installed;
448
449 static void remove_tempfile(void)
450 {
451 int i;
452 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
453 if (diff_temp[i].name == diff_temp[i].tmp_path)
454 unlink_or_warn(diff_temp[i].name);
455 diff_temp[i].name = NULL;
456 }
457 }
458
459 static void remove_tempfile_on_signal(int signo)
460 {
461 remove_tempfile();
462 sigchain_pop(signo);
463 raise(signo);
464 }
465
466 static void print_line_count(FILE *file, int count)
467 {
468 switch (count) {
469 case 0:
470 fprintf(file, "0,0");
471 break;
472 case 1:
473 fprintf(file, "1");
474 break;
475 default:
476 fprintf(file, "1,%d", count);
477 break;
478 }
479 }
480
481 static void emit_rewrite_lines(struct emit_callback *ecb,
482 int prefix, const char *data, int size)
483 {
484 const char *endp = NULL;
485 static const char *nneof = " No newline at end of file\n";
486 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
487 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
488
489 while (0 < size) {
490 int len;
491
492 endp = memchr(data, '\n', size);
493 len = endp ? (endp - data + 1) : size;
494 if (prefix != '+') {
495 ecb->lno_in_preimage++;
496 emit_line_0(ecb->opt, old, reset, '-',
497 data, len);
498 } else {
499 ecb->lno_in_postimage++;
500 emit_add_line(reset, ecb, data, len);
501 }
502 size -= len;
503 data += len;
504 }
505 if (!endp) {
506 const char *plain = diff_get_color(ecb->color_diff,
507 DIFF_PLAIN);
508 emit_line_0(ecb->opt, plain, reset, '\\',
509 nneof, strlen(nneof));
510 }
511 }
512
513 static void emit_rewrite_diff(const char *name_a,
514 const char *name_b,
515 struct diff_filespec *one,
516 struct diff_filespec *two,
517 struct userdiff_driver *textconv_one,
518 struct userdiff_driver *textconv_two,
519 struct diff_options *o)
520 {
521 int lc_a, lc_b;
522 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
523 const char *name_a_tab, *name_b_tab;
524 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
525 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
526 const char *reset = diff_get_color(color_diff, DIFF_RESET);
527 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
528 const char *a_prefix, *b_prefix;
529 char *data_one, *data_two;
530 size_t size_one, size_two;
531 struct emit_callback ecbdata;
532 char *line_prefix = "";
533 struct strbuf *msgbuf;
534
535 if (o && o->output_prefix) {
536 msgbuf = o->output_prefix(o, o->output_prefix_data);
537 line_prefix = msgbuf->buf;
538 }
539
540 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
541 a_prefix = o->b_prefix;
542 b_prefix = o->a_prefix;
543 } else {
544 a_prefix = o->a_prefix;
545 b_prefix = o->b_prefix;
546 }
547
548 name_a += (*name_a == '/');
549 name_b += (*name_b == '/');
550 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
551 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
552
553 strbuf_reset(&a_name);
554 strbuf_reset(&b_name);
555 quote_two_c_style(&a_name, a_prefix, name_a, 0);
556 quote_two_c_style(&b_name, b_prefix, name_b, 0);
557
558 size_one = fill_textconv(textconv_one, one, &data_one);
559 size_two = fill_textconv(textconv_two, two, &data_two);
560
561 memset(&ecbdata, 0, sizeof(ecbdata));
562 ecbdata.color_diff = color_diff;
563 ecbdata.found_changesp = &o->found_changes;
564 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
565 ecbdata.opt = o;
566 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
567 mmfile_t mf1, mf2;
568 mf1.ptr = (char *)data_one;
569 mf2.ptr = (char *)data_two;
570 mf1.size = size_one;
571 mf2.size = size_two;
572 check_blank_at_eof(&mf1, &mf2, &ecbdata);
573 }
574 ecbdata.lno_in_preimage = 1;
575 ecbdata.lno_in_postimage = 1;
576
577 lc_a = count_lines(data_one, size_one);
578 lc_b = count_lines(data_two, size_two);
579 fprintf(o->file,
580 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
581 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
582 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
583 line_prefix, fraginfo);
584 if (!o->irreversible_delete)
585 print_line_count(o->file, lc_a);
586 else
587 fprintf(o->file, "?,?");
588 fprintf(o->file, " +");
589 print_line_count(o->file, lc_b);
590 fprintf(o->file, " @@%s\n", reset);
591 if (lc_a && !o->irreversible_delete)
592 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
593 if (lc_b)
594 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
595 if (textconv_one)
596 free((char *)data_one);
597 if (textconv_two)
598 free((char *)data_two);
599 }
600
601 struct diff_words_buffer {
602 mmfile_t text;
603 long alloc;
604 struct diff_words_orig {
605 const char *begin, *end;
606 } *orig;
607 int orig_nr, orig_alloc;
608 };
609
610 static void diff_words_append(char *line, unsigned long len,
611 struct diff_words_buffer *buffer)
612 {
613 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
614 line++;
615 len--;
616 memcpy(buffer->text.ptr + buffer->text.size, line, len);
617 buffer->text.size += len;
618 buffer->text.ptr[buffer->text.size] = '\0';
619 }
620
621 struct diff_words_style_elem {
622 const char *prefix;
623 const char *suffix;
624 const char *color; /* NULL; filled in by the setup code if
625 * color is enabled */
626 };
627
628 struct diff_words_style {
629 enum diff_words_type type;
630 struct diff_words_style_elem new, old, ctx;
631 const char *newline;
632 };
633
634 static struct diff_words_style diff_words_styles[] = {
635 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
636 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
637 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
638 };
639
640 struct diff_words_data {
641 struct diff_words_buffer minus, plus;
642 const char *current_plus;
643 int last_minus;
644 struct diff_options *opt;
645 regex_t *word_regex;
646 enum diff_words_type type;
647 struct diff_words_style *style;
648 };
649
650 static int fn_out_diff_words_write_helper(FILE *fp,
651 struct diff_words_style_elem *st_el,
652 const char *newline,
653 size_t count, const char *buf,
654 const char *line_prefix)
655 {
656 int print = 0;
657
658 while (count) {
659 char *p = memchr(buf, '\n', count);
660 if (print)
661 fputs(line_prefix, fp);
662 if (p != buf) {
663 if (st_el->color && fputs(st_el->color, fp) < 0)
664 return -1;
665 if (fputs(st_el->prefix, fp) < 0 ||
666 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
667 fputs(st_el->suffix, fp) < 0)
668 return -1;
669 if (st_el->color && *st_el->color
670 && fputs(GIT_COLOR_RESET, fp) < 0)
671 return -1;
672 }
673 if (!p)
674 return 0;
675 if (fputs(newline, fp) < 0)
676 return -1;
677 count -= p + 1 - buf;
678 buf = p + 1;
679 print = 1;
680 }
681 return 0;
682 }
683
684 /*
685 * '--color-words' algorithm can be described as:
686 *
687 * 1. collect a the minus/plus lines of a diff hunk, divided into
688 * minus-lines and plus-lines;
689 *
690 * 2. break both minus-lines and plus-lines into words and
691 * place them into two mmfile_t with one word for each line;
692 *
693 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
694 *
695 * And for the common parts of the both file, we output the plus side text.
696 * diff_words->current_plus is used to trace the current position of the plus file
697 * which printed. diff_words->last_minus is used to trace the last minus word
698 * printed.
699 *
700 * For '--graph' to work with '--color-words', we need to output the graph prefix
701 * on each line of color words output. Generally, there are two conditions on
702 * which we should output the prefix.
703 *
704 * 1. diff_words->last_minus == 0 &&
705 * diff_words->current_plus == diff_words->plus.text.ptr
706 *
707 * that is: the plus text must start as a new line, and if there is no minus
708 * word printed, a graph prefix must be printed.
709 *
710 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
711 * *(diff_words->current_plus - 1) == '\n'
712 *
713 * that is: a graph prefix must be printed following a '\n'
714 */
715 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
716 {
717 if ((diff_words->last_minus == 0 &&
718 diff_words->current_plus == diff_words->plus.text.ptr) ||
719 (diff_words->current_plus > diff_words->plus.text.ptr &&
720 *(diff_words->current_plus - 1) == '\n')) {
721 return 1;
722 } else {
723 return 0;
724 }
725 }
726
727 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
728 {
729 struct diff_words_data *diff_words = priv;
730 struct diff_words_style *style = diff_words->style;
731 int minus_first, minus_len, plus_first, plus_len;
732 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
733 struct diff_options *opt = diff_words->opt;
734 struct strbuf *msgbuf;
735 char *line_prefix = "";
736
737 if (line[0] != '@' || parse_hunk_header(line, len,
738 &minus_first, &minus_len, &plus_first, &plus_len))
739 return;
740
741 assert(opt);
742 if (opt->output_prefix) {
743 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
744 line_prefix = msgbuf->buf;
745 }
746
747 /* POSIX requires that first be decremented by one if len == 0... */
748 if (minus_len) {
749 minus_begin = diff_words->minus.orig[minus_first].begin;
750 minus_end =
751 diff_words->minus.orig[minus_first + minus_len - 1].end;
752 } else
753 minus_begin = minus_end =
754 diff_words->minus.orig[minus_first].end;
755
756 if (plus_len) {
757 plus_begin = diff_words->plus.orig[plus_first].begin;
758 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
759 } else
760 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
761
762 if (color_words_output_graph_prefix(diff_words)) {
763 fputs(line_prefix, diff_words->opt->file);
764 }
765 if (diff_words->current_plus != plus_begin) {
766 fn_out_diff_words_write_helper(diff_words->opt->file,
767 &style->ctx, style->newline,
768 plus_begin - diff_words->current_plus,
769 diff_words->current_plus, line_prefix);
770 if (*(plus_begin - 1) == '\n')
771 fputs(line_prefix, diff_words->opt->file);
772 }
773 if (minus_begin != minus_end) {
774 fn_out_diff_words_write_helper(diff_words->opt->file,
775 &style->old, style->newline,
776 minus_end - minus_begin, minus_begin,
777 line_prefix);
778 }
779 if (plus_begin != plus_end) {
780 fn_out_diff_words_write_helper(diff_words->opt->file,
781 &style->new, style->newline,
782 plus_end - plus_begin, plus_begin,
783 line_prefix);
784 }
785
786 diff_words->current_plus = plus_end;
787 diff_words->last_minus = minus_first;
788 }
789
790 /* This function starts looking at *begin, and returns 0 iff a word was found. */
791 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
792 int *begin, int *end)
793 {
794 if (word_regex && *begin < buffer->size) {
795 regmatch_t match[1];
796 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
797 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
798 '\n', match[0].rm_eo - match[0].rm_so);
799 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
800 *begin += match[0].rm_so;
801 return *begin >= *end;
802 }
803 return -1;
804 }
805
806 /* find the next word */
807 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
808 (*begin)++;
809 if (*begin >= buffer->size)
810 return -1;
811
812 /* find the end of the word */
813 *end = *begin + 1;
814 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
815 (*end)++;
816
817 return 0;
818 }
819
820 /*
821 * This function splits the words in buffer->text, stores the list with
822 * newline separator into out, and saves the offsets of the original words
823 * in buffer->orig.
824 */
825 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
826 regex_t *word_regex)
827 {
828 int i, j;
829 long alloc = 0;
830
831 out->size = 0;
832 out->ptr = NULL;
833
834 /* fake an empty "0th" word */
835 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
836 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
837 buffer->orig_nr = 1;
838
839 for (i = 0; i < buffer->text.size; i++) {
840 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
841 return;
842
843 /* store original boundaries */
844 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
845 buffer->orig_alloc);
846 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
847 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
848 buffer->orig_nr++;
849
850 /* store one word */
851 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
852 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
853 out->ptr[out->size + j - i] = '\n';
854 out->size += j - i + 1;
855
856 i = j - 1;
857 }
858 }
859
860 /* this executes the word diff on the accumulated buffers */
861 static void diff_words_show(struct diff_words_data *diff_words)
862 {
863 xpparam_t xpp;
864 xdemitconf_t xecfg;
865 mmfile_t minus, plus;
866 struct diff_words_style *style = diff_words->style;
867
868 struct diff_options *opt = diff_words->opt;
869 struct strbuf *msgbuf;
870 char *line_prefix = "";
871
872 assert(opt);
873 if (opt->output_prefix) {
874 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
875 line_prefix = msgbuf->buf;
876 }
877
878 /* special case: only removal */
879 if (!diff_words->plus.text.size) {
880 fputs(line_prefix, diff_words->opt->file);
881 fn_out_diff_words_write_helper(diff_words->opt->file,
882 &style->old, style->newline,
883 diff_words->minus.text.size,
884 diff_words->minus.text.ptr, line_prefix);
885 diff_words->minus.text.size = 0;
886 return;
887 }
888
889 diff_words->current_plus = diff_words->plus.text.ptr;
890 diff_words->last_minus = 0;
891
892 memset(&xpp, 0, sizeof(xpp));
893 memset(&xecfg, 0, sizeof(xecfg));
894 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
895 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
896 xpp.flags = 0;
897 /* as only the hunk header will be parsed, we need a 0-context */
898 xecfg.ctxlen = 0;
899 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
900 &xpp, &xecfg);
901 free(minus.ptr);
902 free(plus.ptr);
903 if (diff_words->current_plus != diff_words->plus.text.ptr +
904 diff_words->plus.text.size) {
905 if (color_words_output_graph_prefix(diff_words))
906 fputs(line_prefix, diff_words->opt->file);
907 fn_out_diff_words_write_helper(diff_words->opt->file,
908 &style->ctx, style->newline,
909 diff_words->plus.text.ptr + diff_words->plus.text.size
910 - diff_words->current_plus, diff_words->current_plus,
911 line_prefix);
912 }
913 diff_words->minus.text.size = diff_words->plus.text.size = 0;
914 }
915
916 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
917 static void diff_words_flush(struct emit_callback *ecbdata)
918 {
919 if (ecbdata->diff_words->minus.text.size ||
920 ecbdata->diff_words->plus.text.size)
921 diff_words_show(ecbdata->diff_words);
922 }
923
924 static void free_diff_words_data(struct emit_callback *ecbdata)
925 {
926 if (ecbdata->diff_words) {
927 diff_words_flush(ecbdata);
928 free (ecbdata->diff_words->minus.text.ptr);
929 free (ecbdata->diff_words->minus.orig);
930 free (ecbdata->diff_words->plus.text.ptr);
931 free (ecbdata->diff_words->plus.orig);
932 if (ecbdata->diff_words->word_regex) {
933 regfree(ecbdata->diff_words->word_regex);
934 free(ecbdata->diff_words->word_regex);
935 }
936 free(ecbdata->diff_words);
937 ecbdata->diff_words = NULL;
938 }
939 }
940
941 const char *diff_get_color(int diff_use_color, enum color_diff ix)
942 {
943 if (diff_use_color)
944 return diff_colors[ix];
945 return "";
946 }
947
948 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
949 {
950 const char *cp;
951 unsigned long allot;
952 size_t l = len;
953
954 if (ecb->truncate)
955 return ecb->truncate(line, len);
956 cp = line;
957 allot = l;
958 while (0 < l) {
959 (void) utf8_width(&cp, &l);
960 if (!cp)
961 break; /* truncated in the middle? */
962 }
963 return allot - l;
964 }
965
966 static void find_lno(const char *line, struct emit_callback *ecbdata)
967 {
968 const char *p;
969 ecbdata->lno_in_preimage = 0;
970 ecbdata->lno_in_postimage = 0;
971 p = strchr(line, '-');
972 if (!p)
973 return; /* cannot happen */
974 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
975 p = strchr(p, '+');
976 if (!p)
977 return; /* cannot happen */
978 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
979 }
980
981 static void fn_out_consume(void *priv, char *line, unsigned long len)
982 {
983 struct emit_callback *ecbdata = priv;
984 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
985 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
986 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
987 struct diff_options *o = ecbdata->opt;
988 char *line_prefix = "";
989 struct strbuf *msgbuf;
990
991 if (o && o->output_prefix) {
992 msgbuf = o->output_prefix(o, o->output_prefix_data);
993 line_prefix = msgbuf->buf;
994 }
995
996 if (ecbdata->header) {
997 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
998 strbuf_reset(ecbdata->header);
999 ecbdata->header = NULL;
1000 }
1001 *(ecbdata->found_changesp) = 1;
1002
1003 if (ecbdata->label_path[0]) {
1004 const char *name_a_tab, *name_b_tab;
1005
1006 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1007 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1008
1009 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1010 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1011 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1012 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1013 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1014 }
1015
1016 if (diff_suppress_blank_empty
1017 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1018 line[0] = '\n';
1019 len = 1;
1020 }
1021
1022 if (line[0] == '@') {
1023 if (ecbdata->diff_words)
1024 diff_words_flush(ecbdata);
1025 len = sane_truncate_line(ecbdata, line, len);
1026 find_lno(line, ecbdata);
1027 emit_hunk_header(ecbdata, line, len);
1028 if (line[len-1] != '\n')
1029 putc('\n', ecbdata->opt->file);
1030 return;
1031 }
1032
1033 if (len < 1) {
1034 emit_line(ecbdata->opt, reset, reset, line, len);
1035 if (ecbdata->diff_words
1036 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1037 fputs("~\n", ecbdata->opt->file);
1038 return;
1039 }
1040
1041 if (ecbdata->diff_words) {
1042 if (line[0] == '-') {
1043 diff_words_append(line, len,
1044 &ecbdata->diff_words->minus);
1045 return;
1046 } else if (line[0] == '+') {
1047 diff_words_append(line, len,
1048 &ecbdata->diff_words->plus);
1049 return;
1050 }
1051 diff_words_flush(ecbdata);
1052 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1053 emit_line(ecbdata->opt, plain, reset, line, len);
1054 fputs("~\n", ecbdata->opt->file);
1055 } else {
1056 /* don't print the prefix character */
1057 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1058 }
1059 return;
1060 }
1061
1062 if (line[0] != '+') {
1063 const char *color =
1064 diff_get_color(ecbdata->color_diff,
1065 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1066 ecbdata->lno_in_preimage++;
1067 if (line[0] == ' ')
1068 ecbdata->lno_in_postimage++;
1069 emit_line(ecbdata->opt, color, reset, line, len);
1070 } else {
1071 ecbdata->lno_in_postimage++;
1072 emit_add_line(reset, ecbdata, line + 1, len - 1);
1073 }
1074 }
1075
1076 static char *pprint_rename(const char *a, const char *b)
1077 {
1078 const char *old = a;
1079 const char *new = b;
1080 struct strbuf name = STRBUF_INIT;
1081 int pfx_length, sfx_length;
1082 int len_a = strlen(a);
1083 int len_b = strlen(b);
1084 int a_midlen, b_midlen;
1085 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1086 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1087
1088 if (qlen_a || qlen_b) {
1089 quote_c_style(a, &name, NULL, 0);
1090 strbuf_addstr(&name, " => ");
1091 quote_c_style(b, &name, NULL, 0);
1092 return strbuf_detach(&name, NULL);
1093 }
1094
1095 /* Find common prefix */
1096 pfx_length = 0;
1097 while (*old && *new && *old == *new) {
1098 if (*old == '/')
1099 pfx_length = old - a + 1;
1100 old++;
1101 new++;
1102 }
1103
1104 /* Find common suffix */
1105 old = a + len_a;
1106 new = b + len_b;
1107 sfx_length = 0;
1108 while (a <= old && b <= new && *old == *new) {
1109 if (*old == '/')
1110 sfx_length = len_a - (old - a);
1111 old--;
1112 new--;
1113 }
1114
1115 /*
1116 * pfx{mid-a => mid-b}sfx
1117 * {pfx-a => pfx-b}sfx
1118 * pfx{sfx-a => sfx-b}
1119 * name-a => name-b
1120 */
1121 a_midlen = len_a - pfx_length - sfx_length;
1122 b_midlen = len_b - pfx_length - sfx_length;
1123 if (a_midlen < 0)
1124 a_midlen = 0;
1125 if (b_midlen < 0)
1126 b_midlen = 0;
1127
1128 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1129 if (pfx_length + sfx_length) {
1130 strbuf_add(&name, a, pfx_length);
1131 strbuf_addch(&name, '{');
1132 }
1133 strbuf_add(&name, a + pfx_length, a_midlen);
1134 strbuf_addstr(&name, " => ");
1135 strbuf_add(&name, b + pfx_length, b_midlen);
1136 if (pfx_length + sfx_length) {
1137 strbuf_addch(&name, '}');
1138 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1139 }
1140 return strbuf_detach(&name, NULL);
1141 }
1142
1143 struct diffstat_t {
1144 int nr;
1145 int alloc;
1146 struct diffstat_file {
1147 char *from_name;
1148 char *name;
1149 char *print_name;
1150 unsigned is_unmerged:1;
1151 unsigned is_binary:1;
1152 unsigned is_renamed:1;
1153 uintmax_t added, deleted;
1154 } **files;
1155 };
1156
1157 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1158 const char *name_a,
1159 const char *name_b)
1160 {
1161 struct diffstat_file *x;
1162 x = xcalloc(sizeof (*x), 1);
1163 if (diffstat->nr == diffstat->alloc) {
1164 diffstat->alloc = alloc_nr(diffstat->alloc);
1165 diffstat->files = xrealloc(diffstat->files,
1166 diffstat->alloc * sizeof(x));
1167 }
1168 diffstat->files[diffstat->nr++] = x;
1169 if (name_b) {
1170 x->from_name = xstrdup(name_a);
1171 x->name = xstrdup(name_b);
1172 x->is_renamed = 1;
1173 }
1174 else {
1175 x->from_name = NULL;
1176 x->name = xstrdup(name_a);
1177 }
1178 return x;
1179 }
1180
1181 static void diffstat_consume(void *priv, char *line, unsigned long len)
1182 {
1183 struct diffstat_t *diffstat = priv;
1184 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1185
1186 if (line[0] == '+')
1187 x->added++;
1188 else if (line[0] == '-')
1189 x->deleted++;
1190 }
1191
1192 const char mime_boundary_leader[] = "------------";
1193
1194 static int scale_linear(int it, int width, int max_change)
1195 {
1196 /*
1197 * make sure that at least one '-' is printed if there were deletions,
1198 * and likewise for '+'.
1199 */
1200 if (max_change < 2)
1201 return it;
1202 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1203 }
1204
1205 static void show_name(FILE *file,
1206 const char *prefix, const char *name, int len)
1207 {
1208 fprintf(file, " %s%-*s |", prefix, len, name);
1209 }
1210
1211 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1212 {
1213 if (cnt <= 0)
1214 return;
1215 fprintf(file, "%s", set);
1216 while (cnt--)
1217 putc(ch, file);
1218 fprintf(file, "%s", reset);
1219 }
1220
1221 static void fill_print_name(struct diffstat_file *file)
1222 {
1223 char *pname;
1224
1225 if (file->print_name)
1226 return;
1227
1228 if (!file->is_renamed) {
1229 struct strbuf buf = STRBUF_INIT;
1230 if (quote_c_style(file->name, &buf, NULL, 0)) {
1231 pname = strbuf_detach(&buf, NULL);
1232 } else {
1233 pname = file->name;
1234 strbuf_release(&buf);
1235 }
1236 } else {
1237 pname = pprint_rename(file->from_name, file->name);
1238 }
1239 file->print_name = pname;
1240 }
1241
1242 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1243 {
1244 int i, len, add, del, adds = 0, dels = 0;
1245 uintmax_t max_change = 0, max_len = 0;
1246 int total_files = data->nr;
1247 int width, name_width;
1248 const char *reset, *add_c, *del_c;
1249 const char *line_prefix = "";
1250 struct strbuf *msg = NULL;
1251
1252 if (data->nr == 0)
1253 return;
1254
1255 if (options->output_prefix) {
1256 msg = options->output_prefix(options, options->output_prefix_data);
1257 line_prefix = msg->buf;
1258 }
1259
1260 width = options->stat_width ? options->stat_width : 80;
1261 name_width = options->stat_name_width ? options->stat_name_width : 50;
1262
1263 /* Sanity: give at least 5 columns to the graph,
1264 * but leave at least 10 columns for the name.
1265 */
1266 if (width < 25)
1267 width = 25;
1268 if (name_width < 10)
1269 name_width = 10;
1270 else if (width < name_width + 15)
1271 name_width = width - 15;
1272
1273 /* Find the longest filename and max number of changes */
1274 reset = diff_get_color_opt(options, DIFF_RESET);
1275 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1276 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1277
1278 for (i = 0; i < data->nr; i++) {
1279 struct diffstat_file *file = data->files[i];
1280 uintmax_t change = file->added + file->deleted;
1281 fill_print_name(file);
1282 len = strlen(file->print_name);
1283 if (max_len < len)
1284 max_len = len;
1285
1286 if (file->is_binary || file->is_unmerged)
1287 continue;
1288 if (max_change < change)
1289 max_change = change;
1290 }
1291
1292 /* Compute the width of the graph part;
1293 * 10 is for one blank at the beginning of the line plus
1294 * " | count " between the name and the graph.
1295 *
1296 * From here on, name_width is the width of the name area,
1297 * and width is the width of the graph area.
1298 */
1299 name_width = (name_width < max_len) ? name_width : max_len;
1300 if (width < (name_width + 10) + max_change)
1301 width = width - (name_width + 10);
1302 else
1303 width = max_change;
1304
1305 for (i = 0; i < data->nr; i++) {
1306 const char *prefix = "";
1307 char *name = data->files[i]->print_name;
1308 uintmax_t added = data->files[i]->added;
1309 uintmax_t deleted = data->files[i]->deleted;
1310 int name_len;
1311
1312 /*
1313 * "scale" the filename
1314 */
1315 len = name_width;
1316 name_len = strlen(name);
1317 if (name_width < name_len) {
1318 char *slash;
1319 prefix = "...";
1320 len -= 3;
1321 name += name_len - len;
1322 slash = strchr(name, '/');
1323 if (slash)
1324 name = slash;
1325 }
1326
1327 if (data->files[i]->is_binary) {
1328 fprintf(options->file, "%s", line_prefix);
1329 show_name(options->file, prefix, name, len);
1330 fprintf(options->file, " Bin ");
1331 fprintf(options->file, "%s%"PRIuMAX"%s",
1332 del_c, deleted, reset);
1333 fprintf(options->file, " -> ");
1334 fprintf(options->file, "%s%"PRIuMAX"%s",
1335 add_c, added, reset);
1336 fprintf(options->file, " bytes");
1337 fprintf(options->file, "\n");
1338 continue;
1339 }
1340 else if (data->files[i]->is_unmerged) {
1341 fprintf(options->file, "%s", line_prefix);
1342 show_name(options->file, prefix, name, len);
1343 fprintf(options->file, " Unmerged\n");
1344 continue;
1345 }
1346 else if (!data->files[i]->is_renamed &&
1347 (added + deleted == 0)) {
1348 total_files--;
1349 continue;
1350 }
1351
1352 /*
1353 * scale the add/delete
1354 */
1355 add = added;
1356 del = deleted;
1357 adds += add;
1358 dels += del;
1359
1360 if (width <= max_change) {
1361 add = scale_linear(add, width, max_change);
1362 del = scale_linear(del, width, max_change);
1363 }
1364 fprintf(options->file, "%s", line_prefix);
1365 show_name(options->file, prefix, name, len);
1366 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1367 added + deleted ? " " : "");
1368 show_graph(options->file, '+', add, add_c, reset);
1369 show_graph(options->file, '-', del, del_c, reset);
1370 fprintf(options->file, "\n");
1371 }
1372 fprintf(options->file, "%s", line_prefix);
1373 fprintf(options->file,
1374 " %d files changed, %d insertions(+), %d deletions(-)\n",
1375 total_files, adds, dels);
1376 }
1377
1378 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1379 {
1380 int i, adds = 0, dels = 0, total_files = data->nr;
1381
1382 if (data->nr == 0)
1383 return;
1384
1385 for (i = 0; i < data->nr; i++) {
1386 if (!data->files[i]->is_binary &&
1387 !data->files[i]->is_unmerged) {
1388 int added = data->files[i]->added;
1389 int deleted= data->files[i]->deleted;
1390 if (!data->files[i]->is_renamed &&
1391 (added + deleted == 0)) {
1392 total_files--;
1393 } else {
1394 adds += added;
1395 dels += deleted;
1396 }
1397 }
1398 }
1399 if (options->output_prefix) {
1400 struct strbuf *msg = NULL;
1401 msg = options->output_prefix(options,
1402 options->output_prefix_data);
1403 fprintf(options->file, "%s", msg->buf);
1404 }
1405 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1406 total_files, adds, dels);
1407 }
1408
1409 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1410 {
1411 int i;
1412
1413 if (data->nr == 0)
1414 return;
1415
1416 for (i = 0; i < data->nr; i++) {
1417 struct diffstat_file *file = data->files[i];
1418
1419 if (options->output_prefix) {
1420 struct strbuf *msg = NULL;
1421 msg = options->output_prefix(options,
1422 options->output_prefix_data);
1423 fprintf(options->file, "%s", msg->buf);
1424 }
1425
1426 if (file->is_binary)
1427 fprintf(options->file, "-\t-\t");
1428 else
1429 fprintf(options->file,
1430 "%"PRIuMAX"\t%"PRIuMAX"\t",
1431 file->added, file->deleted);
1432 if (options->line_termination) {
1433 fill_print_name(file);
1434 if (!file->is_renamed)
1435 write_name_quoted(file->name, options->file,
1436 options->line_termination);
1437 else {
1438 fputs(file->print_name, options->file);
1439 putc(options->line_termination, options->file);
1440 }
1441 } else {
1442 if (file->is_renamed) {
1443 putc('\0', options->file);
1444 write_name_quoted(file->from_name, options->file, '\0');
1445 }
1446 write_name_quoted(file->name, options->file, '\0');
1447 }
1448 }
1449 }
1450
1451 struct dirstat_file {
1452 const char *name;
1453 unsigned long changed;
1454 };
1455
1456 struct dirstat_dir {
1457 struct dirstat_file *files;
1458 int alloc, nr, percent, cumulative;
1459 };
1460
1461 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1462 unsigned long changed, const char *base, int baselen)
1463 {
1464 unsigned long this_dir = 0;
1465 unsigned int sources = 0;
1466 const char *line_prefix = "";
1467 struct strbuf *msg = NULL;
1468
1469 if (opt->output_prefix) {
1470 msg = opt->output_prefix(opt, opt->output_prefix_data);
1471 line_prefix = msg->buf;
1472 }
1473
1474 while (dir->nr) {
1475 struct dirstat_file *f = dir->files;
1476 int namelen = strlen(f->name);
1477 unsigned long this;
1478 char *slash;
1479
1480 if (namelen < baselen)
1481 break;
1482 if (memcmp(f->name, base, baselen))
1483 break;
1484 slash = strchr(f->name + baselen, '/');
1485 if (slash) {
1486 int newbaselen = slash + 1 - f->name;
1487 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1488 sources++;
1489 } else {
1490 this = f->changed;
1491 dir->files++;
1492 dir->nr--;
1493 sources += 2;
1494 }
1495 this_dir += this;
1496 }
1497
1498 /*
1499 * We don't report dirstat's for
1500 * - the top level
1501 * - or cases where everything came from a single directory
1502 * under this directory (sources == 1).
1503 */
1504 if (baselen && sources != 1) {
1505 int permille = this_dir * 1000 / changed;
1506 if (permille) {
1507 int percent = permille / 10;
1508 if (percent >= dir->percent) {
1509 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1510 percent, permille % 10, baselen, base);
1511 if (!dir->cumulative)
1512 return 0;
1513 }
1514 }
1515 }
1516 return this_dir;
1517 }
1518
1519 static int dirstat_compare(const void *_a, const void *_b)
1520 {
1521 const struct dirstat_file *a = _a;
1522 const struct dirstat_file *b = _b;
1523 return strcmp(a->name, b->name);
1524 }
1525
1526 static void show_dirstat(struct diff_options *options)
1527 {
1528 int i;
1529 unsigned long changed;
1530 struct dirstat_dir dir;
1531 struct diff_queue_struct *q = &diff_queued_diff;
1532
1533 dir.files = NULL;
1534 dir.alloc = 0;
1535 dir.nr = 0;
1536 dir.percent = options->dirstat_percent;
1537 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1538
1539 changed = 0;
1540 for (i = 0; i < q->nr; i++) {
1541 struct diff_filepair *p = q->queue[i];
1542 const char *name;
1543 unsigned long copied, added, damage;
1544
1545 name = p->one->path ? p->one->path : p->two->path;
1546
1547 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1548 diff_populate_filespec(p->one, 0);
1549 diff_populate_filespec(p->two, 0);
1550 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1551 &copied, &added);
1552 diff_free_filespec_data(p->one);
1553 diff_free_filespec_data(p->two);
1554 } else if (DIFF_FILE_VALID(p->one)) {
1555 diff_populate_filespec(p->one, 1);
1556 copied = added = 0;
1557 diff_free_filespec_data(p->one);
1558 } else if (DIFF_FILE_VALID(p->two)) {
1559 diff_populate_filespec(p->two, 1);
1560 copied = 0;
1561 added = p->two->size;
1562 diff_free_filespec_data(p->two);
1563 } else
1564 continue;
1565
1566 /*
1567 * Original minus copied is the removed material,
1568 * added is the new material. They are both damages
1569 * made to the preimage. In --dirstat-by-file mode, count
1570 * damaged files, not damaged lines. This is done by
1571 * counting only a single damaged line per file.
1572 */
1573 damage = (p->one->size - copied) + added;
1574 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE) && damage > 0)
1575 damage = 1;
1576
1577 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1578 dir.files[dir.nr].name = name;
1579 dir.files[dir.nr].changed = damage;
1580 changed += damage;
1581 dir.nr++;
1582 }
1583
1584 /* This can happen even with many files, if everything was renames */
1585 if (!changed)
1586 return;
1587
1588 /* Show all directories with more than x% of the changes */
1589 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1590 gather_dirstat(options, &dir, changed, "", 0);
1591 }
1592
1593 static void free_diffstat_info(struct diffstat_t *diffstat)
1594 {
1595 int i;
1596 for (i = 0; i < diffstat->nr; i++) {
1597 struct diffstat_file *f = diffstat->files[i];
1598 if (f->name != f->print_name)
1599 free(f->print_name);
1600 free(f->name);
1601 free(f->from_name);
1602 free(f);
1603 }
1604 free(diffstat->files);
1605 }
1606
1607 struct checkdiff_t {
1608 const char *filename;
1609 int lineno;
1610 int conflict_marker_size;
1611 struct diff_options *o;
1612 unsigned ws_rule;
1613 unsigned status;
1614 };
1615
1616 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1617 {
1618 char firstchar;
1619 int cnt;
1620
1621 if (len < marker_size + 1)
1622 return 0;
1623 firstchar = line[0];
1624 switch (firstchar) {
1625 case '=': case '>': case '<': case '|':
1626 break;
1627 default:
1628 return 0;
1629 }
1630 for (cnt = 1; cnt < marker_size; cnt++)
1631 if (line[cnt] != firstchar)
1632 return 0;
1633 /* line[1] thru line[marker_size-1] are same as firstchar */
1634 if (len < marker_size + 1 || !isspace(line[marker_size]))
1635 return 0;
1636 return 1;
1637 }
1638
1639 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1640 {
1641 struct checkdiff_t *data = priv;
1642 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1643 int marker_size = data->conflict_marker_size;
1644 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1645 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1646 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1647 char *err;
1648 char *line_prefix = "";
1649 struct strbuf *msgbuf;
1650
1651 assert(data->o);
1652 if (data->o->output_prefix) {
1653 msgbuf = data->o->output_prefix(data->o,
1654 data->o->output_prefix_data);
1655 line_prefix = msgbuf->buf;
1656 }
1657
1658 if (line[0] == '+') {
1659 unsigned bad;
1660 data->lineno++;
1661 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1662 data->status |= 1;
1663 fprintf(data->o->file,
1664 "%s%s:%d: leftover conflict marker\n",
1665 line_prefix, data->filename, data->lineno);
1666 }
1667 bad = ws_check(line + 1, len - 1, data->ws_rule);
1668 if (!bad)
1669 return;
1670 data->status |= bad;
1671 err = whitespace_error_string(bad);
1672 fprintf(data->o->file, "%s%s:%d: %s.\n",
1673 line_prefix, data->filename, data->lineno, err);
1674 free(err);
1675 emit_line(data->o, set, reset, line, 1);
1676 ws_check_emit(line + 1, len - 1, data->ws_rule,
1677 data->o->file, set, reset, ws);
1678 } else if (line[0] == ' ') {
1679 data->lineno++;
1680 } else if (line[0] == '@') {
1681 char *plus = strchr(line, '+');
1682 if (plus)
1683 data->lineno = strtol(plus, NULL, 10) - 1;
1684 else
1685 die("invalid diff");
1686 }
1687 }
1688
1689 static unsigned char *deflate_it(char *data,
1690 unsigned long size,
1691 unsigned long *result_size)
1692 {
1693 int bound;
1694 unsigned char *deflated;
1695 z_stream stream;
1696
1697 memset(&stream, 0, sizeof(stream));
1698 deflateInit(&stream, zlib_compression_level);
1699 bound = deflateBound(&stream, size);
1700 deflated = xmalloc(bound);
1701 stream.next_out = deflated;
1702 stream.avail_out = bound;
1703
1704 stream.next_in = (unsigned char *)data;
1705 stream.avail_in = size;
1706 while (deflate(&stream, Z_FINISH) == Z_OK)
1707 ; /* nothing */
1708 deflateEnd(&stream);
1709 *result_size = stream.total_out;
1710 return deflated;
1711 }
1712
1713 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1714 {
1715 void *cp;
1716 void *delta;
1717 void *deflated;
1718 void *data;
1719 unsigned long orig_size;
1720 unsigned long delta_size;
1721 unsigned long deflate_size;
1722 unsigned long data_size;
1723
1724 /* We could do deflated delta, or we could do just deflated two,
1725 * whichever is smaller.
1726 */
1727 delta = NULL;
1728 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1729 if (one->size && two->size) {
1730 delta = diff_delta(one->ptr, one->size,
1731 two->ptr, two->size,
1732 &delta_size, deflate_size);
1733 if (delta) {
1734 void *to_free = delta;
1735 orig_size = delta_size;
1736 delta = deflate_it(delta, delta_size, &delta_size);
1737 free(to_free);
1738 }
1739 }
1740
1741 if (delta && delta_size < deflate_size) {
1742 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1743 free(deflated);
1744 data = delta;
1745 data_size = delta_size;
1746 }
1747 else {
1748 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1749 free(delta);
1750 data = deflated;
1751 data_size = deflate_size;
1752 }
1753
1754 /* emit data encoded in base85 */
1755 cp = data;
1756 while (data_size) {
1757 int bytes = (52 < data_size) ? 52 : data_size;
1758 char line[70];
1759 data_size -= bytes;
1760 if (bytes <= 26)
1761 line[0] = bytes + 'A' - 1;
1762 else
1763 line[0] = bytes - 26 + 'a' - 1;
1764 encode_85(line + 1, cp, bytes);
1765 cp = (char *) cp + bytes;
1766 fprintf(file, "%s", prefix);
1767 fputs(line, file);
1768 fputc('\n', file);
1769 }
1770 fprintf(file, "%s\n", prefix);
1771 free(data);
1772 }
1773
1774 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1775 {
1776 fprintf(file, "%sGIT binary patch\n", prefix);
1777 emit_binary_diff_body(file, one, two, prefix);
1778 emit_binary_diff_body(file, two, one, prefix);
1779 }
1780
1781 static void diff_filespec_load_driver(struct diff_filespec *one)
1782 {
1783 /* Use already-loaded driver */
1784 if (one->driver)
1785 return;
1786
1787 if (S_ISREG(one->mode))
1788 one->driver = userdiff_find_by_path(one->path);
1789
1790 /* Fallback to default settings */
1791 if (!one->driver)
1792 one->driver = userdiff_find_by_name("default");
1793 }
1794
1795 int diff_filespec_is_binary(struct diff_filespec *one)
1796 {
1797 if (one->is_binary == -1) {
1798 diff_filespec_load_driver(one);
1799 if (one->driver->binary != -1)
1800 one->is_binary = one->driver->binary;
1801 else {
1802 if (!one->data && DIFF_FILE_VALID(one))
1803 diff_populate_filespec(one, 0);
1804 if (one->data)
1805 one->is_binary = buffer_is_binary(one->data,
1806 one->size);
1807 if (one->is_binary == -1)
1808 one->is_binary = 0;
1809 }
1810 }
1811 return one->is_binary;
1812 }
1813
1814 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1815 {
1816 diff_filespec_load_driver(one);
1817 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1818 }
1819
1820 static const char *userdiff_word_regex(struct diff_filespec *one)
1821 {
1822 diff_filespec_load_driver(one);
1823 return one->driver->word_regex;
1824 }
1825
1826 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1827 {
1828 if (!options->a_prefix)
1829 options->a_prefix = a;
1830 if (!options->b_prefix)
1831 options->b_prefix = b;
1832 }
1833
1834 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1835 {
1836 if (!DIFF_FILE_VALID(one))
1837 return NULL;
1838
1839 diff_filespec_load_driver(one);
1840 if (!one->driver->textconv)
1841 return NULL;
1842
1843 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1844 struct notes_cache *c = xmalloc(sizeof(*c));
1845 struct strbuf name = STRBUF_INIT;
1846
1847 strbuf_addf(&name, "textconv/%s", one->driver->name);
1848 notes_cache_init(c, name.buf, one->driver->textconv);
1849 one->driver->textconv_cache = c;
1850 }
1851
1852 return one->driver;
1853 }
1854
1855 static void builtin_diff(const char *name_a,
1856 const char *name_b,
1857 struct diff_filespec *one,
1858 struct diff_filespec *two,
1859 const char *xfrm_msg,
1860 int must_show_header,
1861 struct diff_options *o,
1862 int complete_rewrite)
1863 {
1864 mmfile_t mf1, mf2;
1865 const char *lbl[2];
1866 char *a_one, *b_two;
1867 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1868 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1869 const char *a_prefix, *b_prefix;
1870 struct userdiff_driver *textconv_one = NULL;
1871 struct userdiff_driver *textconv_two = NULL;
1872 struct strbuf header = STRBUF_INIT;
1873 struct strbuf *msgbuf;
1874 char *line_prefix = "";
1875
1876 if (o->output_prefix) {
1877 msgbuf = o->output_prefix(o, o->output_prefix_data);
1878 line_prefix = msgbuf->buf;
1879 }
1880
1881 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1882 (!one->mode || S_ISGITLINK(one->mode)) &&
1883 (!two->mode || S_ISGITLINK(two->mode))) {
1884 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1885 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1886 show_submodule_summary(o->file, one ? one->path : two->path,
1887 one->sha1, two->sha1, two->dirty_submodule,
1888 del, add, reset);
1889 return;
1890 }
1891
1892 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1893 textconv_one = get_textconv(one);
1894 textconv_two = get_textconv(two);
1895 }
1896
1897 diff_set_mnemonic_prefix(o, "a/", "b/");
1898 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1899 a_prefix = o->b_prefix;
1900 b_prefix = o->a_prefix;
1901 } else {
1902 a_prefix = o->a_prefix;
1903 b_prefix = o->b_prefix;
1904 }
1905
1906 /* Never use a non-valid filename anywhere if at all possible */
1907 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1908 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1909
1910 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1911 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1912 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1913 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1914 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1915 if (lbl[0][0] == '/') {
1916 /* /dev/null */
1917 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1918 if (xfrm_msg)
1919 strbuf_addstr(&header, xfrm_msg);
1920 must_show_header = 1;
1921 }
1922 else if (lbl[1][0] == '/') {
1923 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1924 if (xfrm_msg)
1925 strbuf_addstr(&header, xfrm_msg);
1926 must_show_header = 1;
1927 }
1928 else {
1929 if (one->mode != two->mode) {
1930 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1931 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1932 must_show_header = 1;
1933 }
1934 if (xfrm_msg)
1935 strbuf_addstr(&header, xfrm_msg);
1936
1937 /*
1938 * we do not run diff between different kind
1939 * of objects.
1940 */
1941 if ((one->mode ^ two->mode) & S_IFMT)
1942 goto free_ab_and_return;
1943 if (complete_rewrite &&
1944 (textconv_one || !diff_filespec_is_binary(one)) &&
1945 (textconv_two || !diff_filespec_is_binary(two))) {
1946 fprintf(o->file, "%s", header.buf);
1947 strbuf_reset(&header);
1948 emit_rewrite_diff(name_a, name_b, one, two,
1949 textconv_one, textconv_two, o);
1950 o->found_changes = 1;
1951 goto free_ab_and_return;
1952 }
1953 }
1954
1955 if (o->irreversible_delete && lbl[1][0] == '/') {
1956 fprintf(o->file, "%s", header.buf);
1957 strbuf_reset(&header);
1958 goto free_ab_and_return;
1959 } else if (!DIFF_OPT_TST(o, TEXT) &&
1960 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1961 (!textconv_two && diff_filespec_is_binary(two)) )) {
1962 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1963 die("unable to read files to diff");
1964 /* Quite common confusing case */
1965 if (mf1.size == mf2.size &&
1966 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
1967 if (must_show_header)
1968 fprintf(o->file, "%s", header.buf);
1969 goto free_ab_and_return;
1970 }
1971 fprintf(o->file, "%s", header.buf);
1972 strbuf_reset(&header);
1973 if (DIFF_OPT_TST(o, BINARY))
1974 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
1975 else
1976 fprintf(o->file, "%sBinary files %s and %s differ\n",
1977 line_prefix, lbl[0], lbl[1]);
1978 o->found_changes = 1;
1979 } else {
1980 /* Crazy xdl interfaces.. */
1981 const char *diffopts = getenv("GIT_DIFF_OPTS");
1982 xpparam_t xpp;
1983 xdemitconf_t xecfg;
1984 struct emit_callback ecbdata;
1985 const struct userdiff_funcname *pe;
1986
1987 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
1988 fprintf(o->file, "%s", header.buf);
1989 strbuf_reset(&header);
1990 }
1991
1992 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
1993 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
1994
1995 pe = diff_funcname_pattern(one);
1996 if (!pe)
1997 pe = diff_funcname_pattern(two);
1998
1999 memset(&xpp, 0, sizeof(xpp));
2000 memset(&xecfg, 0, sizeof(xecfg));
2001 memset(&ecbdata, 0, sizeof(ecbdata));
2002 ecbdata.label_path = lbl;
2003 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2004 ecbdata.found_changesp = &o->found_changes;
2005 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2006 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2007 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2008 ecbdata.opt = o;
2009 ecbdata.header = header.len ? &header : NULL;
2010 xpp.flags = o->xdl_opts;
2011 xecfg.ctxlen = o->context;
2012 xecfg.interhunkctxlen = o->interhunkcontext;
2013 xecfg.flags = XDL_EMIT_FUNCNAMES;
2014 if (pe)
2015 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2016 if (!diffopts)
2017 ;
2018 else if (!prefixcmp(diffopts, "--unified="))
2019 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2020 else if (!prefixcmp(diffopts, "-u"))
2021 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2022 if (o->word_diff) {
2023 int i;
2024
2025 ecbdata.diff_words =
2026 xcalloc(1, sizeof(struct diff_words_data));
2027 ecbdata.diff_words->type = o->word_diff;
2028 ecbdata.diff_words->opt = o;
2029 if (!o->word_regex)
2030 o->word_regex = userdiff_word_regex(one);
2031 if (!o->word_regex)
2032 o->word_regex = userdiff_word_regex(two);
2033 if (!o->word_regex)
2034 o->word_regex = diff_word_regex_cfg;
2035 if (o->word_regex) {
2036 ecbdata.diff_words->word_regex = (regex_t *)
2037 xmalloc(sizeof(regex_t));
2038 if (regcomp(ecbdata.diff_words->word_regex,
2039 o->word_regex,
2040 REG_EXTENDED | REG_NEWLINE))
2041 die ("Invalid regular expression: %s",
2042 o->word_regex);
2043 }
2044 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2045 if (o->word_diff == diff_words_styles[i].type) {
2046 ecbdata.diff_words->style =
2047 &diff_words_styles[i];
2048 break;
2049 }
2050 }
2051 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2052 struct diff_words_style *st = ecbdata.diff_words->style;
2053 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2054 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2055 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2056 }
2057 }
2058 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2059 &xpp, &xecfg);
2060 if (o->word_diff)
2061 free_diff_words_data(&ecbdata);
2062 if (textconv_one)
2063 free(mf1.ptr);
2064 if (textconv_two)
2065 free(mf2.ptr);
2066 xdiff_clear_find_func(&xecfg);
2067 }
2068
2069 free_ab_and_return:
2070 strbuf_release(&header);
2071 diff_free_filespec_data(one);
2072 diff_free_filespec_data(two);
2073 free(a_one);
2074 free(b_two);
2075 return;
2076 }
2077
2078 static void builtin_diffstat(const char *name_a, const char *name_b,
2079 struct diff_filespec *one,
2080 struct diff_filespec *two,
2081 struct diffstat_t *diffstat,
2082 struct diff_options *o,
2083 int complete_rewrite)
2084 {
2085 mmfile_t mf1, mf2;
2086 struct diffstat_file *data;
2087
2088 data = diffstat_add(diffstat, name_a, name_b);
2089
2090 if (!one || !two) {
2091 data->is_unmerged = 1;
2092 return;
2093 }
2094
2095 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2096 data->is_binary = 1;
2097 data->added = diff_filespec_size(two);
2098 data->deleted = diff_filespec_size(one);
2099 }
2100
2101 else if (complete_rewrite) {
2102 diff_populate_filespec(one, 0);
2103 diff_populate_filespec(two, 0);
2104 data->deleted = count_lines(one->data, one->size);
2105 data->added = count_lines(two->data, two->size);
2106 }
2107
2108 else {
2109 /* Crazy xdl interfaces.. */
2110 xpparam_t xpp;
2111 xdemitconf_t xecfg;
2112
2113 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2114 die("unable to read files to diff");
2115
2116 memset(&xpp, 0, sizeof(xpp));
2117 memset(&xecfg, 0, sizeof(xecfg));
2118 xpp.flags = o->xdl_opts;
2119 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2120 &xpp, &xecfg);
2121 }
2122
2123 diff_free_filespec_data(one);
2124 diff_free_filespec_data(two);
2125 }
2126
2127 static void builtin_checkdiff(const char *name_a, const char *name_b,
2128 const char *attr_path,
2129 struct diff_filespec *one,
2130 struct diff_filespec *two,
2131 struct diff_options *o)
2132 {
2133 mmfile_t mf1, mf2;
2134 struct checkdiff_t data;
2135
2136 if (!two)
2137 return;
2138
2139 memset(&data, 0, sizeof(data));
2140 data.filename = name_b ? name_b : name_a;
2141 data.lineno = 0;
2142 data.o = o;
2143 data.ws_rule = whitespace_rule(attr_path);
2144 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2145
2146 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2147 die("unable to read files to diff");
2148
2149 /*
2150 * All the other codepaths check both sides, but not checking
2151 * the "old" side here is deliberate. We are checking the newly
2152 * introduced changes, and as long as the "new" side is text, we
2153 * can and should check what it introduces.
2154 */
2155 if (diff_filespec_is_binary(two))
2156 goto free_and_return;
2157 else {
2158 /* Crazy xdl interfaces.. */
2159 xpparam_t xpp;
2160 xdemitconf_t xecfg;
2161
2162 memset(&xpp, 0, sizeof(xpp));
2163 memset(&xecfg, 0, sizeof(xecfg));
2164 xecfg.ctxlen = 1; /* at least one context line */
2165 xpp.flags = 0;
2166 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2167 &xpp, &xecfg);
2168
2169 if (data.ws_rule & WS_BLANK_AT_EOF) {
2170 struct emit_callback ecbdata;
2171 int blank_at_eof;
2172
2173 ecbdata.ws_rule = data.ws_rule;
2174 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2175 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2176
2177 if (blank_at_eof) {
2178 static char *err;
2179 if (!err)
2180 err = whitespace_error_string(WS_BLANK_AT_EOF);
2181 fprintf(o->file, "%s:%d: %s.\n",
2182 data.filename, blank_at_eof, err);
2183 data.status = 1; /* report errors */
2184 }
2185 }
2186 }
2187 free_and_return:
2188 diff_free_filespec_data(one);
2189 diff_free_filespec_data(two);
2190 if (data.status)
2191 DIFF_OPT_SET(o, CHECK_FAILED);
2192 }
2193
2194 struct diff_filespec *alloc_filespec(const char *path)
2195 {
2196 int namelen = strlen(path);
2197 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2198
2199 memset(spec, 0, sizeof(*spec));
2200 spec->path = (char *)(spec + 1);
2201 memcpy(spec->path, path, namelen+1);
2202 spec->count = 1;
2203 spec->is_binary = -1;
2204 return spec;
2205 }
2206
2207 void free_filespec(struct diff_filespec *spec)
2208 {
2209 if (!--spec->count) {
2210 diff_free_filespec_data(spec);
2211 free(spec);
2212 }
2213 }
2214
2215 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2216 unsigned short mode)
2217 {
2218 if (mode) {
2219 spec->mode = canon_mode(mode);
2220 hashcpy(spec->sha1, sha1);
2221 spec->sha1_valid = !is_null_sha1(sha1);
2222 }
2223 }
2224
2225 /*
2226 * Given a name and sha1 pair, if the index tells us the file in
2227 * the work tree has that object contents, return true, so that
2228 * prepare_temp_file() does not have to inflate and extract.
2229 */
2230 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2231 {
2232 struct cache_entry *ce;
2233 struct stat st;
2234 int pos, len;
2235
2236 /*
2237 * We do not read the cache ourselves here, because the
2238 * benchmark with my previous version that always reads cache
2239 * shows that it makes things worse for diff-tree comparing
2240 * two linux-2.6 kernel trees in an already checked out work
2241 * tree. This is because most diff-tree comparisons deal with
2242 * only a small number of files, while reading the cache is
2243 * expensive for a large project, and its cost outweighs the
2244 * savings we get by not inflating the object to a temporary
2245 * file. Practically, this code only helps when we are used
2246 * by diff-cache --cached, which does read the cache before
2247 * calling us.
2248 */
2249 if (!active_cache)
2250 return 0;
2251
2252 /* We want to avoid the working directory if our caller
2253 * doesn't need the data in a normal file, this system
2254 * is rather slow with its stat/open/mmap/close syscalls,
2255 * and the object is contained in a pack file. The pack
2256 * is probably already open and will be faster to obtain
2257 * the data through than the working directory. Loose
2258 * objects however would tend to be slower as they need
2259 * to be individually opened and inflated.
2260 */
2261 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2262 return 0;
2263
2264 len = strlen(name);
2265 pos = cache_name_pos(name, len);
2266 if (pos < 0)
2267 return 0;
2268 ce = active_cache[pos];
2269
2270 /*
2271 * This is not the sha1 we are looking for, or
2272 * unreusable because it is not a regular file.
2273 */
2274 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2275 return 0;
2276
2277 /*
2278 * If ce is marked as "assume unchanged", there is no
2279 * guarantee that work tree matches what we are looking for.
2280 */
2281 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2282 return 0;
2283
2284 /*
2285 * If ce matches the file in the work tree, we can reuse it.
2286 */
2287 if (ce_uptodate(ce) ||
2288 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2289 return 1;
2290
2291 return 0;
2292 }
2293
2294 static int populate_from_stdin(struct diff_filespec *s)
2295 {
2296 struct strbuf buf = STRBUF_INIT;
2297 size_t size = 0;
2298
2299 if (strbuf_read(&buf, 0, 0) < 0)
2300 return error("error while reading from stdin %s",
2301 strerror(errno));
2302
2303 s->should_munmap = 0;
2304 s->data = strbuf_detach(&buf, &size);
2305 s->size = size;
2306 s->should_free = 1;
2307 return 0;
2308 }
2309
2310 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2311 {
2312 int len;
2313 char *data = xmalloc(100), *dirty = "";
2314
2315 /* Are we looking at the work tree? */
2316 if (s->dirty_submodule)
2317 dirty = "-dirty";
2318
2319 len = snprintf(data, 100,
2320 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2321 s->data = data;
2322 s->size = len;
2323 s->should_free = 1;
2324 if (size_only) {
2325 s->data = NULL;
2326 free(data);
2327 }
2328 return 0;
2329 }
2330
2331 /*
2332 * While doing rename detection and pickaxe operation, we may need to
2333 * grab the data for the blob (or file) for our own in-core comparison.
2334 * diff_filespec has data and size fields for this purpose.
2335 */
2336 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2337 {
2338 int err = 0;
2339 if (!DIFF_FILE_VALID(s))
2340 die("internal error: asking to populate invalid file.");
2341 if (S_ISDIR(s->mode))
2342 return -1;
2343
2344 if (s->data)
2345 return 0;
2346
2347 if (size_only && 0 < s->size)
2348 return 0;
2349
2350 if (S_ISGITLINK(s->mode))
2351 return diff_populate_gitlink(s, size_only);
2352
2353 if (!s->sha1_valid ||
2354 reuse_worktree_file(s->path, s->sha1, 0)) {
2355 struct strbuf buf = STRBUF_INIT;
2356 struct stat st;
2357 int fd;
2358
2359 if (!strcmp(s->path, "-"))
2360 return populate_from_stdin(s);
2361
2362 if (lstat(s->path, &st) < 0) {
2363 if (errno == ENOENT) {
2364 err_empty:
2365 err = -1;
2366 empty:
2367 s->data = (char *)"";
2368 s->size = 0;
2369 return err;
2370 }
2371 }
2372 s->size = xsize_t(st.st_size);
2373 if (!s->size)
2374 goto empty;
2375 if (S_ISLNK(st.st_mode)) {
2376 struct strbuf sb = STRBUF_INIT;
2377
2378 if (strbuf_readlink(&sb, s->path, s->size))
2379 goto err_empty;
2380 s->size = sb.len;
2381 s->data = strbuf_detach(&sb, NULL);
2382 s->should_free = 1;
2383 return 0;
2384 }
2385 if (size_only)
2386 return 0;
2387 fd = open(s->path, O_RDONLY);
2388 if (fd < 0)
2389 goto err_empty;
2390 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2391 close(fd);
2392 s->should_munmap = 1;
2393
2394 /*
2395 * Convert from working tree format to canonical git format
2396 */
2397 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2398 size_t size = 0;
2399 munmap(s->data, s->size);
2400 s->should_munmap = 0;
2401 s->data = strbuf_detach(&buf, &size);
2402 s->size = size;
2403 s->should_free = 1;
2404 }
2405 }
2406 else {
2407 enum object_type type;
2408 if (size_only) {
2409 type = sha1_object_info(s->sha1, &s->size);
2410 if (type < 0)
2411 die("unable to read %s", sha1_to_hex(s->sha1));
2412 } else {
2413 s->data = read_sha1_file(s->sha1, &type, &s->size);
2414 if (!s->data)
2415 die("unable to read %s", sha1_to_hex(s->sha1));
2416 s->should_free = 1;
2417 }
2418 }
2419 return 0;
2420 }
2421
2422 void diff_free_filespec_blob(struct diff_filespec *s)
2423 {
2424 if (s->should_free)
2425 free(s->data);
2426 else if (s->should_munmap)
2427 munmap(s->data, s->size);
2428
2429 if (s->should_free || s->should_munmap) {
2430 s->should_free = s->should_munmap = 0;
2431 s->data = NULL;
2432 }
2433 }
2434
2435 void diff_free_filespec_data(struct diff_filespec *s)
2436 {
2437 diff_free_filespec_blob(s);
2438 free(s->cnt_data);
2439 s->cnt_data = NULL;
2440 }
2441
2442 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2443 void *blob,
2444 unsigned long size,
2445 const unsigned char *sha1,
2446 int mode)
2447 {
2448 int fd;
2449 struct strbuf buf = STRBUF_INIT;
2450 struct strbuf template = STRBUF_INIT;
2451 char *path_dup = xstrdup(path);
2452 const char *base = basename(path_dup);
2453
2454 /* Generate "XXXXXX_basename.ext" */
2455 strbuf_addstr(&template, "XXXXXX_");
2456 strbuf_addstr(&template, base);
2457
2458 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2459 strlen(base) + 1);
2460 if (fd < 0)
2461 die_errno("unable to create temp-file");
2462 if (convert_to_working_tree(path,
2463 (const char *)blob, (size_t)size, &buf)) {
2464 blob = buf.buf;
2465 size = buf.len;
2466 }
2467 if (write_in_full(fd, blob, size) != size)
2468 die_errno("unable to write temp-file");
2469 close(fd);
2470 temp->name = temp->tmp_path;
2471 strcpy(temp->hex, sha1_to_hex(sha1));
2472 temp->hex[40] = 0;
2473 sprintf(temp->mode, "%06o", mode);
2474 strbuf_release(&buf);
2475 strbuf_release(&template);
2476 free(path_dup);
2477 }
2478
2479 static struct diff_tempfile *prepare_temp_file(const char *name,
2480 struct diff_filespec *one)
2481 {
2482 struct diff_tempfile *temp = claim_diff_tempfile();
2483
2484 if (!DIFF_FILE_VALID(one)) {
2485 not_a_valid_file:
2486 /* A '-' entry produces this for file-2, and
2487 * a '+' entry produces this for file-1.
2488 */
2489 temp->name = "/dev/null";
2490 strcpy(temp->hex, ".");
2491 strcpy(temp->mode, ".");
2492 return temp;
2493 }
2494
2495 if (!remove_tempfile_installed) {
2496 atexit(remove_tempfile);
2497 sigchain_push_common(remove_tempfile_on_signal);
2498 remove_tempfile_installed = 1;
2499 }
2500
2501 if (!one->sha1_valid ||
2502 reuse_worktree_file(name, one->sha1, 1)) {
2503 struct stat st;
2504 if (lstat(name, &st) < 0) {
2505 if (errno == ENOENT)
2506 goto not_a_valid_file;
2507 die_errno("stat(%s)", name);
2508 }
2509 if (S_ISLNK(st.st_mode)) {
2510 struct strbuf sb = STRBUF_INIT;
2511 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2512 die_errno("readlink(%s)", name);
2513 prep_temp_blob(name, temp, sb.buf, sb.len,
2514 (one->sha1_valid ?
2515 one->sha1 : null_sha1),
2516 (one->sha1_valid ?
2517 one->mode : S_IFLNK));
2518 strbuf_release(&sb);
2519 }
2520 else {
2521 /* we can borrow from the file in the work tree */
2522 temp->name = name;
2523 if (!one->sha1_valid)
2524 strcpy(temp->hex, sha1_to_hex(null_sha1));
2525 else
2526 strcpy(temp->hex, sha1_to_hex(one->sha1));
2527 /* Even though we may sometimes borrow the
2528 * contents from the work tree, we always want
2529 * one->mode. mode is trustworthy even when
2530 * !(one->sha1_valid), as long as
2531 * DIFF_FILE_VALID(one).
2532 */
2533 sprintf(temp->mode, "%06o", one->mode);
2534 }
2535 return temp;
2536 }
2537 else {
2538 if (diff_populate_filespec(one, 0))
2539 die("cannot read data blob for %s", one->path);
2540 prep_temp_blob(name, temp, one->data, one->size,
2541 one->sha1, one->mode);
2542 }
2543 return temp;
2544 }
2545
2546 /* An external diff command takes:
2547 *
2548 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2549 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2550 *
2551 */
2552 static void run_external_diff(const char *pgm,
2553 const char *name,
2554 const char *other,
2555 struct diff_filespec *one,
2556 struct diff_filespec *two,
2557 const char *xfrm_msg,
2558 int complete_rewrite)
2559 {
2560 const char *spawn_arg[10];
2561 int retval;
2562 const char **arg = &spawn_arg[0];
2563
2564 if (one && two) {
2565 struct diff_tempfile *temp_one, *temp_two;
2566 const char *othername = (other ? other : name);
2567 temp_one = prepare_temp_file(name, one);
2568 temp_two = prepare_temp_file(othername, two);
2569 *arg++ = pgm;
2570 *arg++ = name;
2571 *arg++ = temp_one->name;
2572 *arg++ = temp_one->hex;
2573 *arg++ = temp_one->mode;
2574 *arg++ = temp_two->name;
2575 *arg++ = temp_two->hex;
2576 *arg++ = temp_two->mode;
2577 if (other) {
2578 *arg++ = other;
2579 *arg++ = xfrm_msg;
2580 }
2581 } else {
2582 *arg++ = pgm;
2583 *arg++ = name;
2584 }
2585 *arg = NULL;
2586 fflush(NULL);
2587 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2588 remove_tempfile();
2589 if (retval) {
2590 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2591 exit(1);
2592 }
2593 }
2594
2595 static int similarity_index(struct diff_filepair *p)
2596 {
2597 return p->score * 100 / MAX_SCORE;
2598 }
2599
2600 static void fill_metainfo(struct strbuf *msg,
2601 const char *name,
2602 const char *other,
2603 struct diff_filespec *one,
2604 struct diff_filespec *two,
2605 struct diff_options *o,
2606 struct diff_filepair *p,
2607 int *must_show_header,
2608 int use_color)
2609 {
2610 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2611 const char *reset = diff_get_color(use_color, DIFF_RESET);
2612 struct strbuf *msgbuf;
2613 char *line_prefix = "";
2614
2615 *must_show_header = 1;
2616 if (o->output_prefix) {
2617 msgbuf = o->output_prefix(o, o->output_prefix_data);
2618 line_prefix = msgbuf->buf;
2619 }
2620 strbuf_init(msg, PATH_MAX * 2 + 300);
2621 switch (p->status) {
2622 case DIFF_STATUS_COPIED:
2623 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2624 line_prefix, set, similarity_index(p));
2625 strbuf_addf(msg, "%s\n%s%scopy from ",
2626 reset, line_prefix, set);
2627 quote_c_style(name, msg, NULL, 0);
2628 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2629 quote_c_style(other, msg, NULL, 0);
2630 strbuf_addf(msg, "%s\n", reset);
2631 break;
2632 case DIFF_STATUS_RENAMED:
2633 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2634 line_prefix, set, similarity_index(p));
2635 strbuf_addf(msg, "%s\n%s%srename from ",
2636 reset, line_prefix, set);
2637 quote_c_style(name, msg, NULL, 0);
2638 strbuf_addf(msg, "%s\n%s%srename to ",
2639 reset, line_prefix, set);
2640 quote_c_style(other, msg, NULL, 0);
2641 strbuf_addf(msg, "%s\n", reset);
2642 break;
2643 case DIFF_STATUS_MODIFIED:
2644 if (p->score) {
2645 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2646 line_prefix,
2647 set, similarity_index(p), reset);
2648 break;
2649 }
2650 /* fallthru */
2651 default:
2652 *must_show_header = 0;
2653 }
2654 if (one && two && hashcmp(one->sha1, two->sha1)) {
2655 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2656
2657 if (DIFF_OPT_TST(o, BINARY)) {
2658 mmfile_t mf;
2659 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2660 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2661 abbrev = 40;
2662 }
2663 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2664 find_unique_abbrev(one->sha1, abbrev));
2665 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2666 if (one->mode == two->mode)
2667 strbuf_addf(msg, " %06o", one->mode);
2668 strbuf_addf(msg, "%s\n", reset);
2669 }
2670 }
2671
2672 static void run_diff_cmd(const char *pgm,
2673 const char *name,
2674 const char *other,
2675 const char *attr_path,
2676 struct diff_filespec *one,
2677 struct diff_filespec *two,
2678 struct strbuf *msg,
2679 struct diff_options *o,
2680 struct diff_filepair *p)
2681 {
2682 const char *xfrm_msg = NULL;
2683 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2684 int must_show_header = 0;
2685
2686 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2687 pgm = NULL;
2688 else {
2689 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2690 if (drv && drv->external)
2691 pgm = drv->external;
2692 }
2693
2694 if (msg) {
2695 /*
2696 * don't use colors when the header is intended for an
2697 * external diff driver
2698 */
2699 fill_metainfo(msg, name, other, one, two, o, p,
2700 &must_show_header,
2701 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2702 xfrm_msg = msg->len ? msg->buf : NULL;
2703 }
2704
2705 if (pgm) {
2706 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2707 complete_rewrite);
2708 return;
2709 }
2710 if (one && two)
2711 builtin_diff(name, other ? other : name,
2712 one, two, xfrm_msg, must_show_header,
2713 o, complete_rewrite);
2714 else
2715 fprintf(o->file, "* Unmerged path %s\n", name);
2716 }
2717
2718 static void diff_fill_sha1_info(struct diff_filespec *one)
2719 {
2720 if (DIFF_FILE_VALID(one)) {
2721 if (!one->sha1_valid) {
2722 struct stat st;
2723 if (!strcmp(one->path, "-")) {
2724 hashcpy(one->sha1, null_sha1);
2725 return;
2726 }
2727 if (lstat(one->path, &st) < 0)
2728 die_errno("stat '%s'", one->path);
2729 if (index_path(one->sha1, one->path, &st, 0))
2730 die("cannot hash %s", one->path);
2731 }
2732 }
2733 else
2734 hashclr(one->sha1);
2735 }
2736
2737 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2738 {
2739 /* Strip the prefix but do not molest /dev/null and absolute paths */
2740 if (*namep && **namep != '/') {
2741 *namep += prefix_length;
2742 if (**namep == '/')
2743 ++*namep;
2744 }
2745 if (*otherp && **otherp != '/') {
2746 *otherp += prefix_length;
2747 if (**otherp == '/')
2748 ++*otherp;
2749 }
2750 }
2751
2752 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2753 {
2754 const char *pgm = external_diff();
2755 struct strbuf msg;
2756 struct diff_filespec *one = p->one;
2757 struct diff_filespec *two = p->two;
2758 const char *name;
2759 const char *other;
2760 const char *attr_path;
2761
2762 name = p->one->path;
2763 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2764 attr_path = name;
2765 if (o->prefix_length)
2766 strip_prefix(o->prefix_length, &name, &other);
2767
2768 if (DIFF_PAIR_UNMERGED(p)) {
2769 run_diff_cmd(pgm, name, NULL, attr_path,
2770 NULL, NULL, NULL, o, p);
2771 return;
2772 }
2773
2774 diff_fill_sha1_info(one);
2775 diff_fill_sha1_info(two);
2776
2777 if (!pgm &&
2778 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2779 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2780 /*
2781 * a filepair that changes between file and symlink
2782 * needs to be split into deletion and creation.
2783 */
2784 struct diff_filespec *null = alloc_filespec(two->path);
2785 run_diff_cmd(NULL, name, other, attr_path,
2786 one, null, &msg, o, p);
2787 free(null);
2788 strbuf_release(&msg);
2789
2790 null = alloc_filespec(one->path);
2791 run_diff_cmd(NULL, name, other, attr_path,
2792 null, two, &msg, o, p);
2793 free(null);
2794 }
2795 else
2796 run_diff_cmd(pgm, name, other, attr_path,
2797 one, two, &msg, o, p);
2798
2799 strbuf_release(&msg);
2800 }
2801
2802 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2803 struct diffstat_t *diffstat)
2804 {
2805 const char *name;
2806 const char *other;
2807 int complete_rewrite = 0;
2808
2809 if (DIFF_PAIR_UNMERGED(p)) {
2810 /* unmerged */
2811 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2812 return;
2813 }
2814
2815 name = p->one->path;
2816 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2817
2818 if (o->prefix_length)
2819 strip_prefix(o->prefix_length, &name, &other);
2820
2821 diff_fill_sha1_info(p->one);
2822 diff_fill_sha1_info(p->two);
2823
2824 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2825 complete_rewrite = 1;
2826 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2827 }
2828
2829 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2830 {
2831 const char *name;
2832 const char *other;
2833 const char *attr_path;
2834
2835 if (DIFF_PAIR_UNMERGED(p)) {
2836 /* unmerged */
2837 return;
2838 }
2839
2840 name = p->one->path;
2841 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2842 attr_path = other ? other : name;
2843
2844 if (o->prefix_length)
2845 strip_prefix(o->prefix_length, &name, &other);
2846
2847 diff_fill_sha1_info(p->one);
2848 diff_fill_sha1_info(p->two);
2849
2850 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2851 }
2852
2853 void diff_setup(struct diff_options *options)
2854 {
2855 memcpy(options, &default_diff_options, sizeof(*options));
2856
2857 options->file = stdout;
2858
2859 options->line_termination = '\n';
2860 options->break_opt = -1;
2861 options->rename_limit = -1;
2862 options->dirstat_percent = 3;
2863 options->context = 3;
2864
2865 options->change = diff_change;
2866 options->add_remove = diff_addremove;
2867 if (diff_use_color_default > 0)
2868 DIFF_OPT_SET(options, COLOR_DIFF);
2869 options->detect_rename = diff_detect_rename_default;
2870
2871 if (diff_no_prefix) {
2872 options->a_prefix = options->b_prefix = "";
2873 } else if (!diff_mnemonic_prefix) {
2874 options->a_prefix = "a/";
2875 options->b_prefix = "b/";
2876 }
2877 }
2878
2879 int diff_setup_done(struct diff_options *options)
2880 {
2881 int count = 0;
2882
2883 if (options->output_format & DIFF_FORMAT_NAME)
2884 count++;
2885 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2886 count++;
2887 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2888 count++;
2889 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2890 count++;
2891 if (count > 1)
2892 die("--name-only, --name-status, --check and -s are mutually exclusive");
2893
2894 /*
2895 * Most of the time we can say "there are changes"
2896 * only by checking if there are changed paths, but
2897 * --ignore-whitespace* options force us to look
2898 * inside contents.
2899 */
2900
2901 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2902 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2903 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2904 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2905 else
2906 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2907
2908 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2909 options->detect_rename = DIFF_DETECT_COPY;
2910
2911 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2912 options->prefix = NULL;
2913 if (options->prefix)
2914 options->prefix_length = strlen(options->prefix);
2915 else
2916 options->prefix_length = 0;
2917
2918 if (options->output_format & (DIFF_FORMAT_NAME |
2919 DIFF_FORMAT_NAME_STATUS |
2920 DIFF_FORMAT_CHECKDIFF |
2921 DIFF_FORMAT_NO_OUTPUT))
2922 options->output_format &= ~(DIFF_FORMAT_RAW |
2923 DIFF_FORMAT_NUMSTAT |
2924 DIFF_FORMAT_DIFFSTAT |
2925 DIFF_FORMAT_SHORTSTAT |
2926 DIFF_FORMAT_DIRSTAT |
2927 DIFF_FORMAT_SUMMARY |
2928 DIFF_FORMAT_PATCH);
2929
2930 /*
2931 * These cases always need recursive; we do not drop caller-supplied
2932 * recursive bits for other formats here.
2933 */
2934 if (options->output_format & (DIFF_FORMAT_PATCH |
2935 DIFF_FORMAT_NUMSTAT |
2936 DIFF_FORMAT_DIFFSTAT |
2937 DIFF_FORMAT_SHORTSTAT |
2938 DIFF_FORMAT_DIRSTAT |
2939 DIFF_FORMAT_SUMMARY |
2940 DIFF_FORMAT_CHECKDIFF))
2941 DIFF_OPT_SET(options, RECURSIVE);
2942 /*
2943 * Also pickaxe would not work very well if you do not say recursive
2944 */
2945 if (options->pickaxe)
2946 DIFF_OPT_SET(options, RECURSIVE);
2947 /*
2948 * When patches are generated, submodules diffed against the work tree
2949 * must be checked for dirtiness too so it can be shown in the output
2950 */
2951 if (options->output_format & DIFF_FORMAT_PATCH)
2952 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2953
2954 if (options->detect_rename && options->rename_limit < 0)
2955 options->rename_limit = diff_rename_limit_default;
2956 if (options->setup & DIFF_SETUP_USE_CACHE) {
2957 if (!active_cache)
2958 /* read-cache does not die even when it fails
2959 * so it is safe for us to do this here. Also
2960 * it does not smudge active_cache or active_nr
2961 * when it fails, so we do not have to worry about
2962 * cleaning it up ourselves either.
2963 */
2964 read_cache();
2965 }
2966 if (options->abbrev <= 0 || 40 < options->abbrev)
2967 options->abbrev = 40; /* full */
2968
2969 /*
2970 * It does not make sense to show the first hit we happened
2971 * to have found. It does not make sense not to return with
2972 * exit code in such a case either.
2973 */
2974 if (DIFF_OPT_TST(options, QUICK)) {
2975 options->output_format = DIFF_FORMAT_NO_OUTPUT;
2976 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
2977 }
2978
2979 return 0;
2980 }
2981
2982 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
2983 {
2984 char c, *eq;
2985 int len;
2986
2987 if (*arg != '-')
2988 return 0;
2989 c = *++arg;
2990 if (!c)
2991 return 0;
2992 if (c == arg_short) {
2993 c = *++arg;
2994 if (!c)
2995 return 1;
2996 if (val && isdigit(c)) {
2997 char *end;
2998 int n = strtoul(arg, &end, 10);
2999 if (*end)
3000 return 0;
3001 *val = n;
3002 return 1;
3003 }
3004 return 0;
3005 }
3006 if (c != '-')
3007 return 0;
3008 arg++;
3009 eq = strchr(arg, '=');
3010 if (eq)
3011 len = eq - arg;
3012 else
3013 len = strlen(arg);
3014 if (!len || strncmp(arg, arg_long, len))
3015 return 0;
3016 if (eq) {
3017 int n;
3018 char *end;
3019 if (!isdigit(*++eq))
3020 return 0;
3021 n = strtoul(eq, &end, 10);
3022 if (*end)
3023 return 0;
3024 *val = n;
3025 }
3026 return 1;
3027 }
3028
3029 static int diff_scoreopt_parse(const char *opt);
3030
3031 static inline int short_opt(char opt, const char **argv,
3032 const char **optarg)
3033 {
3034 const char *arg = argv[0];
3035 if (arg[0] != '-' || arg[1] != opt)
3036 return 0;
3037 if (arg[2] != '\0') {
3038 *optarg = arg + 2;
3039 return 1;
3040 }
3041 if (!argv[1])
3042 die("Option '%c' requires a value", opt);
3043 *optarg = argv[1];
3044 return 2;
3045 }
3046
3047 int parse_long_opt(const char *opt, const char **argv,
3048 const char **optarg)
3049 {
3050 const char *arg = argv[0];
3051 if (arg[0] != '-' || arg[1] != '-')
3052 return 0;
3053 arg += strlen("--");
3054 if (prefixcmp(arg, opt))
3055 return 0;
3056 arg += strlen(opt);
3057 if (*arg == '=') { /* sticked form: --option=value */
3058 *optarg = arg + 1;
3059 return 1;
3060 }
3061 if (*arg != '\0')
3062 return 0;
3063 /* separate form: --option value */
3064 if (!argv[1])
3065 die("Option '--%s' requires a value", opt);
3066 *optarg = argv[1];
3067 return 2;
3068 }
3069
3070 static int stat_opt(struct diff_options *options, const char **av)
3071 {
3072 const char *arg = av[0];
3073 char *end;
3074 int width = options->stat_width;
3075 int name_width = options->stat_name_width;
3076 int argcount = 1;
3077
3078 arg += strlen("--stat");
3079 end = (char *)arg;
3080
3081 switch (*arg) {
3082 case '-':
3083 if (!prefixcmp(arg, "-width")) {
3084 arg += strlen("-width");
3085 if (*arg == '=')
3086 width = strtoul(arg + 1, &end, 10);
3087 else if (!*arg && !av[1])
3088 die("Option '--stat-width' requires a value");
3089 else if (!*arg) {
3090 width = strtoul(av[1], &end, 10);
3091 argcount = 2;
3092 }
3093 } else if (!prefixcmp(arg, "-name-width")) {
3094 arg += strlen("-name-width");
3095 if (*arg == '=')
3096 name_width = strtoul(arg + 1, &end, 10);
3097 else if (!*arg && !av[1])
3098 die("Option '--stat-name-width' requires a value");
3099 else if (!*arg) {
3100 name_width = strtoul(av[1], &end, 10);
3101 argcount = 2;
3102 }
3103 }
3104 break;
3105 case '=':
3106 width = strtoul(arg+1, &end, 10);
3107 if (*end == ',')
3108 name_width = strtoul(end+1, &end, 10);
3109 }
3110
3111 /* Important! This checks all the error cases! */
3112 if (*end)
3113 return 0;
3114 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3115 options->stat_name_width = name_width;
3116 options->stat_width = width;
3117 return argcount;
3118 }
3119
3120 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3121 {
3122 const char *arg = av[0];
3123 const char *optarg;
3124 int argcount;
3125
3126 /* Output format options */
3127 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3128 options->output_format |= DIFF_FORMAT_PATCH;
3129 else if (opt_arg(arg, 'U', "unified", &options->context))
3130 options->output_format |= DIFF_FORMAT_PATCH;
3131 else if (!strcmp(arg, "--raw"))
3132 options->output_format |= DIFF_FORMAT_RAW;
3133 else if (!strcmp(arg, "--patch-with-raw"))
3134 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3135 else if (!strcmp(arg, "--numstat"))
3136 options->output_format |= DIFF_FORMAT_NUMSTAT;
3137 else if (!strcmp(arg, "--shortstat"))
3138 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3139 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3140 options->output_format |= DIFF_FORMAT_DIRSTAT;
3141 else if (!strcmp(arg, "--cumulative")) {
3142 options->output_format |= DIFF_FORMAT_DIRSTAT;
3143 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3144 } else if (opt_arg(arg, 0, "dirstat-by-file",
3145 &options->dirstat_percent)) {
3146 options->output_format |= DIFF_FORMAT_DIRSTAT;
3147 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3148 }
3149 else if (!strcmp(arg, "--check"))
3150 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3151 else if (!strcmp(arg, "--summary"))
3152 options->output_format |= DIFF_FORMAT_SUMMARY;
3153 else if (!strcmp(arg, "--patch-with-stat"))
3154 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3155 else if (!strcmp(arg, "--name-only"))
3156 options->output_format |= DIFF_FORMAT_NAME;
3157 else if (!strcmp(arg, "--name-status"))
3158 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3159 else if (!strcmp(arg, "-s"))
3160 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3161 else if (!prefixcmp(arg, "--stat"))
3162 /* --stat, --stat-width, or --stat-name-width */
3163 return stat_opt(options, av);
3164
3165 /* renames options */
3166 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3167 !strcmp(arg, "--break-rewrites")) {
3168 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3169 return error("invalid argument to -B: %s", arg+2);
3170 }
3171 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3172 !strcmp(arg, "--find-renames")) {
3173 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3174 return error("invalid argument to -M: %s", arg+2);
3175 options->detect_rename = DIFF_DETECT_RENAME;
3176 }
3177 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3178 options->irreversible_delete = 1;
3179 }
3180 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3181 !strcmp(arg, "--find-copies")) {
3182 if (options->detect_rename == DIFF_DETECT_COPY)
3183 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3184 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3185 return error("invalid argument to -C: %s", arg+2);
3186 options->detect_rename = DIFF_DETECT_COPY;
3187 }
3188 else if (!strcmp(arg, "--no-renames"))
3189 options->detect_rename = 0;
3190 else if (!strcmp(arg, "--relative"))
3191 DIFF_OPT_SET(options, RELATIVE_NAME);
3192 else if (!prefixcmp(arg, "--relative=")) {
3193 DIFF_OPT_SET(options, RELATIVE_NAME);
3194 options->prefix = arg + 11;
3195 }
3196
3197 /* xdiff options */
3198 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3199 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3200 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3201 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3202 else if (!strcmp(arg, "--ignore-space-at-eol"))
3203 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3204 else if (!strcmp(arg, "--patience"))
3205 DIFF_XDL_SET(options, PATIENCE_DIFF);
3206
3207 /* flags options */
3208 else if (!strcmp(arg, "--binary")) {
3209 options->output_format |= DIFF_FORMAT_PATCH;
3210 DIFF_OPT_SET(options, BINARY);
3211 }
3212 else if (!strcmp(arg, "--full-index"))
3213 DIFF_OPT_SET(options, FULL_INDEX);
3214 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3215 DIFF_OPT_SET(options, TEXT);
3216 else if (!strcmp(arg, "-R"))
3217 DIFF_OPT_SET(options, REVERSE_DIFF);
3218 else if (!strcmp(arg, "--find-copies-harder"))
3219 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3220 else if (!strcmp(arg, "--follow"))
3221 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3222 else if (!strcmp(arg, "--color"))
3223 DIFF_OPT_SET(options, COLOR_DIFF);
3224 else if (!prefixcmp(arg, "--color=")) {
3225 int value = git_config_colorbool(NULL, arg+8, -1);
3226 if (value == 0)
3227 DIFF_OPT_CLR(options, COLOR_DIFF);
3228 else if (value > 0)
3229 DIFF_OPT_SET(options, COLOR_DIFF);
3230 else
3231 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3232 }
3233 else if (!strcmp(arg, "--no-color"))
3234 DIFF_OPT_CLR(options, COLOR_DIFF);
3235 else if (!strcmp(arg, "--color-words")) {
3236 DIFF_OPT_SET(options, COLOR_DIFF);
3237 options->word_diff = DIFF_WORDS_COLOR;
3238 }
3239 else if (!prefixcmp(arg, "--color-words=")) {
3240 DIFF_OPT_SET(options, COLOR_DIFF);
3241 options->word_diff = DIFF_WORDS_COLOR;
3242 options->word_regex = arg + 14;
3243 }
3244 else if (!strcmp(arg, "--word-diff")) {
3245 if (options->word_diff == DIFF_WORDS_NONE)
3246 options->word_diff = DIFF_WORDS_PLAIN;
3247 }
3248 else if (!prefixcmp(arg, "--word-diff=")) {
3249 const char *type = arg + 12;
3250 if (!strcmp(type, "plain"))
3251 options->word_diff = DIFF_WORDS_PLAIN;
3252 else if (!strcmp(type, "color")) {
3253 DIFF_OPT_SET(options, COLOR_DIFF);
3254 options->word_diff = DIFF_WORDS_COLOR;
3255 }
3256 else if (!strcmp(type, "porcelain"))
3257 options->word_diff = DIFF_WORDS_PORCELAIN;
3258 else if (!strcmp(type, "none"))
3259 options->word_diff = DIFF_WORDS_NONE;
3260 else
3261 die("bad --word-diff argument: %s", type);
3262 }
3263 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3264 if (options->word_diff == DIFF_WORDS_NONE)
3265 options->word_diff = DIFF_WORDS_PLAIN;
3266 options->word_regex = optarg;
3267 return argcount;
3268 }
3269 else if (!strcmp(arg, "--exit-code"))
3270 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3271 else if (!strcmp(arg, "--quiet"))
3272 DIFF_OPT_SET(options, QUICK);
3273 else if (!strcmp(arg, "--ext-diff"))
3274 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3275 else if (!strcmp(arg, "--no-ext-diff"))
3276 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3277 else if (!strcmp(arg, "--textconv"))
3278 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3279 else if (!strcmp(arg, "--no-textconv"))
3280 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3281 else if (!strcmp(arg, "--ignore-submodules")) {
3282 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3283 handle_ignore_submodules_arg(options, "all");
3284 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3285 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3286 handle_ignore_submodules_arg(options, arg + 20);
3287 } else if (!strcmp(arg, "--submodule"))
3288 DIFF_OPT_SET(options, SUBMODULE_LOG);
3289 else if (!prefixcmp(arg, "--submodule=")) {
3290 if (!strcmp(arg + 12, "log"))
3291 DIFF_OPT_SET(options, SUBMODULE_LOG);
3292 }
3293
3294 /* misc options */
3295 else if (!strcmp(arg, "-z"))
3296 options->line_termination = 0;
3297 else if ((argcount = short_opt('l', av, &optarg))) {
3298 options->rename_limit = strtoul(optarg, NULL, 10);
3299 return argcount;
3300 }
3301 else if ((argcount = short_opt('S', av, &optarg))) {
3302 options->pickaxe = optarg;
3303 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3304 return argcount;
3305 } else if ((argcount = short_opt('G', av, &optarg))) {
3306 options->pickaxe = optarg;
3307 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3308 return argcount;
3309 }
3310 else if (!strcmp(arg, "--pickaxe-all"))
3311 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3312 else if (!strcmp(arg, "--pickaxe-regex"))
3313 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3314 else if ((argcount = short_opt('O', av, &optarg))) {
3315 options->orderfile = optarg;
3316 return argcount;
3317 }
3318 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3319 options->filter = optarg;
3320 return argcount;
3321 }
3322 else if (!strcmp(arg, "--abbrev"))
3323 options->abbrev = DEFAULT_ABBREV;
3324 else if (!prefixcmp(arg, "--abbrev=")) {
3325 options->abbrev = strtoul(arg + 9, NULL, 10);
3326 if (options->abbrev < MINIMUM_ABBREV)
3327 options->abbrev = MINIMUM_ABBREV;
3328 else if (40 < options->abbrev)
3329 options->abbrev = 40;
3330 }
3331 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3332 options->a_prefix = optarg;
3333 return argcount;
3334 }
3335 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3336 options->b_prefix = optarg;
3337 return argcount;
3338 }
3339 else if (!strcmp(arg, "--no-prefix"))
3340 options->a_prefix = options->b_prefix = "";
3341 else if (opt_arg(arg, '\0', "inter-hunk-context",
3342 &options->interhunkcontext))
3343 ;
3344 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3345 options->file = fopen(optarg, "w");
3346 if (!options->file)
3347 die_errno("Could not open '%s'", optarg);
3348 options->close_file = 1;
3349 return argcount;
3350 } else
3351 return 0;
3352 return 1;
3353 }
3354
3355 int parse_rename_score(const char **cp_p)
3356 {
3357 unsigned long num, scale;
3358 int ch, dot;
3359 const char *cp = *cp_p;
3360
3361 num = 0;
3362 scale = 1;
3363 dot = 0;
3364 for (;;) {
3365 ch = *cp;
3366 if ( !dot && ch == '.' ) {
3367 scale = 1;
3368 dot = 1;
3369 } else if ( ch == '%' ) {
3370 scale = dot ? scale*100 : 100;
3371 cp++; /* % is always at the end */
3372 break;
3373 } else if ( ch >= '0' && ch <= '9' ) {
3374 if ( scale < 100000 ) {
3375 scale *= 10;
3376 num = (num*10) + (ch-'0');
3377 }
3378 } else {
3379 break;
3380 }
3381 cp++;
3382 }
3383 *cp_p = cp;
3384
3385 /* user says num divided by scale and we say internally that
3386 * is MAX_SCORE * num / scale.
3387 */
3388 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3389 }
3390
3391 static int diff_scoreopt_parse(const char *opt)
3392 {
3393 int opt1, opt2, cmd;
3394
3395 if (*opt++ != '-')
3396 return -1;
3397 cmd = *opt++;
3398 if (cmd == '-') {
3399 /* convert the long-form arguments into short-form versions */
3400 if (!prefixcmp(opt, "break-rewrites")) {
3401 opt += strlen("break-rewrites");
3402 if (*opt == 0 || *opt++ == '=')
3403 cmd = 'B';
3404 } else if (!prefixcmp(opt, "find-copies")) {
3405 opt += strlen("find-copies");
3406 if (*opt == 0 || *opt++ == '=')
3407 cmd = 'C';
3408 } else if (!prefixcmp(opt, "find-renames")) {
3409 opt += strlen("find-renames");
3410 if (*opt == 0 || *opt++ == '=')
3411 cmd = 'M';
3412 }
3413 }
3414 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3415 return -1; /* that is not a -M, -C nor -B option */
3416
3417 opt1 = parse_rename_score(&opt);
3418 if (cmd != 'B')
3419 opt2 = 0;
3420 else {
3421 if (*opt == 0)
3422 opt2 = 0;
3423 else if (*opt != '/')
3424 return -1; /* we expect -B80/99 or -B80 */
3425 else {
3426 opt++;
3427 opt2 = parse_rename_score(&opt);
3428 }
3429 }
3430 if (*opt != 0)
3431 return -1;
3432 return opt1 | (opt2 << 16);
3433 }
3434
3435 struct diff_queue_struct diff_queued_diff;
3436
3437 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3438 {
3439 if (queue->alloc <= queue->nr) {
3440 queue->alloc = alloc_nr(queue->alloc);
3441 queue->queue = xrealloc(queue->queue,
3442 sizeof(dp) * queue->alloc);
3443 }
3444 queue->queue[queue->nr++] = dp;
3445 }
3446
3447 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3448 struct diff_filespec *one,
3449 struct diff_filespec *two)
3450 {
3451 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3452 dp->one = one;
3453 dp->two = two;
3454 if (queue)
3455 diff_q(queue, dp);
3456 return dp;
3457 }
3458
3459 void diff_free_filepair(struct diff_filepair *p)
3460 {
3461 free_filespec(p->one);
3462 free_filespec(p->two);
3463 free(p);
3464 }
3465
3466 /* This is different from find_unique_abbrev() in that
3467 * it stuffs the result with dots for alignment.
3468 */
3469 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3470 {
3471 int abblen;
3472 const char *abbrev;
3473 if (len == 40)
3474 return sha1_to_hex(sha1);
3475
3476 abbrev = find_unique_abbrev(sha1, len);
3477 abblen = strlen(abbrev);
3478 if (abblen < 37) {
3479 static char hex[41];
3480 if (len < abblen && abblen <= len + 2)
3481 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3482 else
3483 sprintf(hex, "%s...", abbrev);
3484 return hex;
3485 }
3486 return sha1_to_hex(sha1);
3487 }
3488
3489 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3490 {
3491 int line_termination = opt->line_termination;
3492 int inter_name_termination = line_termination ? '\t' : '\0';
3493 if (opt->output_prefix) {
3494 struct strbuf *msg = NULL;
3495 msg = opt->output_prefix(opt, opt->output_prefix_data);
3496 fprintf(opt->file, "%s", msg->buf);
3497 }
3498
3499 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3500 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3501 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3502 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3503 }
3504 if (p->score) {
3505 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3506 inter_name_termination);
3507 } else {
3508 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3509 }
3510
3511 if (p->status == DIFF_STATUS_COPIED ||
3512 p->status == DIFF_STATUS_RENAMED) {
3513 const char *name_a, *name_b;
3514 name_a = p->one->path;
3515 name_b = p->two->path;
3516 strip_prefix(opt->prefix_length, &name_a, &name_b);
3517 write_name_quoted(name_a, opt->file, inter_name_termination);
3518 write_name_quoted(name_b, opt->file, line_termination);
3519 } else {
3520 const char *name_a, *name_b;
3521 name_a = p->one->mode ? p->one->path : p->two->path;
3522 name_b = NULL;
3523 strip_prefix(opt->prefix_length, &name_a, &name_b);
3524 write_name_quoted(name_a, opt->file, line_termination);
3525 }
3526 }
3527
3528 int diff_unmodified_pair(struct diff_filepair *p)
3529 {
3530 /* This function is written stricter than necessary to support
3531 * the currently implemented transformers, but the idea is to
3532 * let transformers to produce diff_filepairs any way they want,
3533 * and filter and clean them up here before producing the output.
3534 */
3535 struct diff_filespec *one = p->one, *two = p->two;
3536
3537 if (DIFF_PAIR_UNMERGED(p))
3538 return 0; /* unmerged is interesting */
3539
3540 /* deletion, addition, mode or type change
3541 * and rename are all interesting.
3542 */
3543 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3544 DIFF_PAIR_MODE_CHANGED(p) ||
3545 strcmp(one->path, two->path))
3546 return 0;
3547
3548 /* both are valid and point at the same path. that is, we are
3549 * dealing with a change.
3550 */
3551 if (one->sha1_valid && two->sha1_valid &&
3552 !hashcmp(one->sha1, two->sha1) &&
3553 !one->dirty_submodule && !two->dirty_submodule)
3554 return 1; /* no change */
3555 if (!one->sha1_valid && !two->sha1_valid)
3556 return 1; /* both look at the same file on the filesystem. */
3557 return 0;
3558 }
3559
3560 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3561 {
3562 if (diff_unmodified_pair(p))
3563 return;
3564
3565 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3566 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3567 return; /* no tree diffs in patch format */
3568
3569 run_diff(p, o);
3570 }
3571
3572 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3573 struct diffstat_t *diffstat)
3574 {
3575 if (diff_unmodified_pair(p))
3576 return;
3577
3578 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3579 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3580 return; /* no useful stat for tree diffs */
3581
3582 run_diffstat(p, o, diffstat);
3583 }
3584
3585 static void diff_flush_checkdiff(struct diff_filepair *p,
3586 struct diff_options *o)
3587 {
3588 if (diff_unmodified_pair(p))
3589 return;
3590
3591 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3592 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3593 return; /* nothing to check in tree diffs */
3594
3595 run_checkdiff(p, o);
3596 }
3597
3598 int diff_queue_is_empty(void)
3599 {
3600 struct diff_queue_struct *q = &diff_queued_diff;
3601 int i;
3602 for (i = 0; i < q->nr; i++)
3603 if (!diff_unmodified_pair(q->queue[i]))
3604 return 0;
3605 return 1;
3606 }
3607
3608 #if DIFF_DEBUG
3609 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3610 {
3611 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3612 x, one ? one : "",
3613 s->path,
3614 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3615 s->mode,
3616 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3617 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3618 x, one ? one : "",
3619 s->size, s->xfrm_flags);
3620 }
3621
3622 void diff_debug_filepair(const struct diff_filepair *p, int i)
3623 {
3624 diff_debug_filespec(p->one, i, "one");
3625 diff_debug_filespec(p->two, i, "two");
3626 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3627 p->score, p->status ? p->status : '?',
3628 p->one->rename_used, p->broken_pair);
3629 }
3630
3631 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3632 {
3633 int i;
3634 if (msg)
3635 fprintf(stderr, "%s\n", msg);
3636 fprintf(stderr, "q->nr = %d\n", q->nr);
3637 for (i = 0; i < q->nr; i++) {
3638 struct diff_filepair *p = q->queue[i];
3639 diff_debug_filepair(p, i);
3640 }
3641 }
3642 #endif
3643
3644 static void diff_resolve_rename_copy(void)
3645 {
3646 int i;
3647 struct diff_filepair *p;
3648 struct diff_queue_struct *q = &diff_queued_diff;
3649
3650 diff_debug_queue("resolve-rename-copy", q);
3651
3652 for (i = 0; i < q->nr; i++) {
3653 p = q->queue[i];
3654 p->status = 0; /* undecided */
3655 if (DIFF_PAIR_UNMERGED(p))
3656 p->status = DIFF_STATUS_UNMERGED;
3657 else if (!DIFF_FILE_VALID(p->one))
3658 p->status = DIFF_STATUS_ADDED;
3659 else if (!DIFF_FILE_VALID(p->two))
3660 p->status = DIFF_STATUS_DELETED;
3661 else if (DIFF_PAIR_TYPE_CHANGED(p))
3662 p->status = DIFF_STATUS_TYPE_CHANGED;
3663
3664 /* from this point on, we are dealing with a pair
3665 * whose both sides are valid and of the same type, i.e.
3666 * either in-place edit or rename/copy edit.
3667 */
3668 else if (DIFF_PAIR_RENAME(p)) {
3669 /*
3670 * A rename might have re-connected a broken
3671 * pair up, causing the pathnames to be the
3672 * same again. If so, that's not a rename at
3673 * all, just a modification..
3674 *
3675 * Otherwise, see if this source was used for
3676 * multiple renames, in which case we decrement
3677 * the count, and call it a copy.
3678 */
3679 if (!strcmp(p->one->path, p->two->path))
3680 p->status = DIFF_STATUS_MODIFIED;
3681 else if (--p->one->rename_used > 0)
3682 p->status = DIFF_STATUS_COPIED;
3683 else
3684 p->status = DIFF_STATUS_RENAMED;
3685 }
3686 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3687 p->one->mode != p->two->mode ||
3688 p->one->dirty_submodule ||
3689 p->two->dirty_submodule ||
3690 is_null_sha1(p->one->sha1))
3691 p->status = DIFF_STATUS_MODIFIED;
3692 else {
3693 /* This is a "no-change" entry and should not
3694 * happen anymore, but prepare for broken callers.
3695 */
3696 error("feeding unmodified %s to diffcore",
3697 p->one->path);
3698 p->status = DIFF_STATUS_UNKNOWN;
3699 }
3700 }
3701 diff_debug_queue("resolve-rename-copy done", q);
3702 }
3703
3704 static int check_pair_status(struct diff_filepair *p)
3705 {
3706 switch (p->status) {
3707 case DIFF_STATUS_UNKNOWN:
3708 return 0;
3709 case 0:
3710 die("internal error in diff-resolve-rename-copy");
3711 default:
3712 return 1;
3713 }
3714 }
3715
3716 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3717 {
3718 int fmt = opt->output_format;
3719
3720 if (fmt & DIFF_FORMAT_CHECKDIFF)
3721 diff_flush_checkdiff(p, opt);
3722 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3723 diff_flush_raw(p, opt);
3724 else if (fmt & DIFF_FORMAT_NAME) {
3725 const char *name_a, *name_b;
3726 name_a = p->two->path;
3727 name_b = NULL;
3728 strip_prefix(opt->prefix_length, &name_a, &name_b);
3729 write_name_quoted(name_a, opt->file, opt->line_termination);
3730 }
3731 }
3732
3733 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3734 {
3735 if (fs->mode)
3736 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3737 else
3738 fprintf(file, " %s ", newdelete);
3739 write_name_quoted(fs->path, file, '\n');
3740 }
3741
3742
3743 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3744 const char *line_prefix)
3745 {
3746 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3747 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3748 p->two->mode, show_name ? ' ' : '\n');
3749 if (show_name) {
3750 write_name_quoted(p->two->path, file, '\n');
3751 }
3752 }
3753 }
3754
3755 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3756 const char *line_prefix)
3757 {
3758 char *names = pprint_rename(p->one->path, p->two->path);
3759
3760 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3761 free(names);
3762 show_mode_change(file, p, 0, line_prefix);
3763 }
3764
3765 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3766 {
3767 FILE *file = opt->file;
3768 char *line_prefix = "";
3769
3770 if (opt->output_prefix) {
3771 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3772 line_prefix = buf->buf;
3773 }
3774
3775 switch(p->status) {
3776 case DIFF_STATUS_DELETED:
3777 fputs(line_prefix, file);
3778 show_file_mode_name(file, "delete", p->one);
3779 break;
3780 case DIFF_STATUS_ADDED:
3781 fputs(line_prefix, file);
3782 show_file_mode_name(file, "create", p->two);
3783 break;
3784 case DIFF_STATUS_COPIED:
3785 fputs(line_prefix, file);
3786 show_rename_copy(file, "copy", p, line_prefix);
3787 break;
3788 case DIFF_STATUS_RENAMED:
3789 fputs(line_prefix, file);
3790 show_rename_copy(file, "rename", p, line_prefix);
3791 break;
3792 default:
3793 if (p->score) {
3794 fprintf(file, "%s rewrite ", line_prefix);
3795 write_name_quoted(p->two->path, file, ' ');
3796 fprintf(file, "(%d%%)\n", similarity_index(p));
3797 }
3798 show_mode_change(file, p, !p->score, line_prefix);
3799 break;
3800 }
3801 }
3802
3803 struct patch_id_t {
3804 git_SHA_CTX *ctx;
3805 int patchlen;
3806 };
3807
3808 static int remove_space(char *line, int len)
3809 {
3810 int i;
3811 char *dst = line;
3812 unsigned char c;
3813
3814 for (i = 0; i < len; i++)
3815 if (!isspace((c = line[i])))
3816 *dst++ = c;
3817
3818 return dst - line;
3819 }
3820
3821 static void patch_id_consume(void *priv, char *line, unsigned long len)
3822 {
3823 struct patch_id_t *data = priv;
3824 int new_len;
3825
3826 /* Ignore line numbers when computing the SHA1 of the patch */
3827 if (!prefixcmp(line, "@@ -"))
3828 return;
3829
3830 new_len = remove_space(line, len);
3831
3832 git_SHA1_Update(data->ctx, line, new_len);
3833 data->patchlen += new_len;
3834 }
3835
3836 /* returns 0 upon success, and writes result into sha1 */
3837 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3838 {
3839 struct diff_queue_struct *q = &diff_queued_diff;
3840 int i;
3841 git_SHA_CTX ctx;
3842 struct patch_id_t data;
3843 char buffer[PATH_MAX * 4 + 20];
3844
3845 git_SHA1_Init(&ctx);
3846 memset(&data, 0, sizeof(struct patch_id_t));
3847 data.ctx = &ctx;
3848
3849 for (i = 0; i < q->nr; i++) {
3850 xpparam_t xpp;
3851 xdemitconf_t xecfg;
3852 mmfile_t mf1, mf2;
3853 struct diff_filepair *p = q->queue[i];
3854 int len1, len2;
3855
3856 memset(&xpp, 0, sizeof(xpp));
3857 memset(&xecfg, 0, sizeof(xecfg));
3858 if (p->status == 0)
3859 return error("internal diff status error");
3860 if (p->status == DIFF_STATUS_UNKNOWN)
3861 continue;
3862 if (diff_unmodified_pair(p))
3863 continue;
3864 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3865 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3866 continue;
3867 if (DIFF_PAIR_UNMERGED(p))
3868 continue;
3869
3870 diff_fill_sha1_info(p->one);
3871 diff_fill_sha1_info(p->two);
3872 if (fill_mmfile(&mf1, p->one) < 0 ||
3873 fill_mmfile(&mf2, p->two) < 0)
3874 return error("unable to read files to diff");
3875
3876 len1 = remove_space(p->one->path, strlen(p->one->path));
3877 len2 = remove_space(p->two->path, strlen(p->two->path));
3878 if (p->one->mode == 0)
3879 len1 = snprintf(buffer, sizeof(buffer),
3880 "diff--gita/%.*sb/%.*s"
3881 "newfilemode%06o"
3882 "---/dev/null"
3883 "+++b/%.*s",
3884 len1, p->one->path,
3885 len2, p->two->path,
3886 p->two->mode,
3887 len2, p->two->path);
3888 else if (p->two->mode == 0)
3889 len1 = snprintf(buffer, sizeof(buffer),
3890 "diff--gita/%.*sb/%.*s"
3891 "deletedfilemode%06o"
3892 "---a/%.*s"
3893 "+++/dev/null",
3894 len1, p->one->path,
3895 len2, p->two->path,
3896 p->one->mode,
3897 len1, p->one->path);
3898 else
3899 len1 = snprintf(buffer, sizeof(buffer),
3900 "diff--gita/%.*sb/%.*s"
3901 "---a/%.*s"
3902 "+++b/%.*s",
3903 len1, p->one->path,
3904 len2, p->two->path,
3905 len1, p->one->path,
3906 len2, p->two->path);
3907 git_SHA1_Update(&ctx, buffer, len1);
3908
3909 if (diff_filespec_is_binary(p->one) ||
3910 diff_filespec_is_binary(p->two)) {
3911 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3912 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3913 continue;
3914 }
3915
3916 xpp.flags = 0;
3917 xecfg.ctxlen = 3;
3918 xecfg.flags = 0;
3919 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3920 &xpp, &xecfg);
3921 }
3922
3923 git_SHA1_Final(sha1, &ctx);
3924 return 0;
3925 }
3926
3927 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3928 {
3929 struct diff_queue_struct *q = &diff_queued_diff;
3930 int i;
3931 int result = diff_get_patch_id(options, sha1);
3932
3933 for (i = 0; i < q->nr; i++)
3934 diff_free_filepair(q->queue[i]);
3935
3936 free(q->queue);
3937 DIFF_QUEUE_CLEAR(q);
3938
3939 return result;
3940 }
3941
3942 static int is_summary_empty(const struct diff_queue_struct *q)
3943 {
3944 int i;
3945
3946 for (i = 0; i < q->nr; i++) {
3947 const struct diff_filepair *p = q->queue[i];
3948
3949 switch (p->status) {
3950 case DIFF_STATUS_DELETED:
3951 case DIFF_STATUS_ADDED:
3952 case DIFF_STATUS_COPIED:
3953 case DIFF_STATUS_RENAMED:
3954 return 0;
3955 default:
3956 if (p->score)
3957 return 0;
3958 if (p->one->mode && p->two->mode &&
3959 p->one->mode != p->two->mode)
3960 return 0;
3961 break;
3962 }
3963 }
3964 return 1;
3965 }
3966
3967 static const char rename_limit_warning[] =
3968 "inexact rename detection was skipped due to too many files.";
3969
3970 static const char degrade_cc_to_c_warning[] =
3971 "only found copies from modified paths due to too many files.";
3972
3973 static const char rename_limit_advice[] =
3974 "you may want to set your %s variable to at least "
3975 "%d and retry the command.";
3976
3977 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
3978 {
3979 if (degraded_cc)
3980 warning(degrade_cc_to_c_warning);
3981 else if (needed)
3982 warning(rename_limit_warning);
3983 else
3984 return;
3985 if (0 < needed && needed < 32767)
3986 warning(rename_limit_advice, varname, needed);
3987 }
3988
3989 void diff_flush(struct diff_options *options)
3990 {
3991 struct diff_queue_struct *q = &diff_queued_diff;
3992 int i, output_format = options->output_format;
3993 int separator = 0;
3994
3995 /*
3996 * Order: raw, stat, summary, patch
3997 * or: name/name-status/checkdiff (other bits clear)
3998 */
3999 if (!q->nr)
4000 goto free_queue;
4001
4002 if (output_format & (DIFF_FORMAT_RAW |
4003 DIFF_FORMAT_NAME |
4004 DIFF_FORMAT_NAME_STATUS |
4005 DIFF_FORMAT_CHECKDIFF)) {
4006 for (i = 0; i < q->nr; i++) {
4007 struct diff_filepair *p = q->queue[i];
4008 if (check_pair_status(p))
4009 flush_one_pair(p, options);
4010 }
4011 separator++;
4012 }
4013
4014 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4015 struct diffstat_t diffstat;
4016
4017 memset(&diffstat, 0, sizeof(struct diffstat_t));
4018 for (i = 0; i < q->nr; i++) {
4019 struct diff_filepair *p = q->queue[i];
4020 if (check_pair_status(p))
4021 diff_flush_stat(p, options, &diffstat);
4022 }
4023 if (output_format & DIFF_FORMAT_NUMSTAT)
4024 show_numstat(&diffstat, options);
4025 if (output_format & DIFF_FORMAT_DIFFSTAT)
4026 show_stats(&diffstat, options);
4027 if (output_format & DIFF_FORMAT_SHORTSTAT)
4028 show_shortstats(&diffstat, options);
4029 free_diffstat_info(&diffstat);
4030 separator++;
4031 }
4032 if (output_format & DIFF_FORMAT_DIRSTAT)
4033 show_dirstat(options);
4034
4035 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4036 for (i = 0; i < q->nr; i++) {
4037 diff_summary(options, q->queue[i]);
4038 }
4039 separator++;
4040 }
4041
4042 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4043 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4044 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4045 /*
4046 * run diff_flush_patch for the exit status. setting
4047 * options->file to /dev/null should be safe, becaue we
4048 * aren't supposed to produce any output anyway.
4049 */
4050 if (options->close_file)
4051 fclose(options->file);
4052 options->file = fopen("/dev/null", "w");
4053 if (!options->file)
4054 die_errno("Could not open /dev/null");
4055 options->close_file = 1;
4056 for (i = 0; i < q->nr; i++) {
4057 struct diff_filepair *p = q->queue[i];
4058 if (check_pair_status(p))
4059 diff_flush_patch(p, options);
4060 if (options->found_changes)
4061 break;
4062 }
4063 }
4064
4065 if (output_format & DIFF_FORMAT_PATCH) {
4066 if (separator) {
4067 putc(options->line_termination, options->file);
4068 if (options->stat_sep) {
4069 /* attach patch instead of inline */
4070 fputs(options->stat_sep, options->file);
4071 }
4072 }
4073
4074 for (i = 0; i < q->nr; i++) {
4075 struct diff_filepair *p = q->queue[i];
4076 if (check_pair_status(p))
4077 diff_flush_patch(p, options);
4078 }
4079 }
4080
4081 if (output_format & DIFF_FORMAT_CALLBACK)
4082 options->format_callback(q, options, options->format_callback_data);
4083
4084 for (i = 0; i < q->nr; i++)
4085 diff_free_filepair(q->queue[i]);
4086 free_queue:
4087 free(q->queue);
4088 DIFF_QUEUE_CLEAR(q);
4089 if (options->close_file)
4090 fclose(options->file);
4091
4092 /*
4093 * Report the content-level differences with HAS_CHANGES;
4094 * diff_addremove/diff_change does not set the bit when
4095 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4096 */
4097 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4098 if (options->found_changes)
4099 DIFF_OPT_SET(options, HAS_CHANGES);
4100 else
4101 DIFF_OPT_CLR(options, HAS_CHANGES);
4102 }
4103 }
4104
4105 static void diffcore_apply_filter(const char *filter)
4106 {
4107 int i;
4108 struct diff_queue_struct *q = &diff_queued_diff;
4109 struct diff_queue_struct outq;
4110 DIFF_QUEUE_CLEAR(&outq);
4111
4112 if (!filter)
4113 return;
4114
4115 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4116 int found;
4117 for (i = found = 0; !found && i < q->nr; i++) {
4118 struct diff_filepair *p = q->queue[i];
4119 if (((p->status == DIFF_STATUS_MODIFIED) &&
4120 ((p->score &&
4121 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4122 (!p->score &&
4123 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4124 ((p->status != DIFF_STATUS_MODIFIED) &&
4125 strchr(filter, p->status)))
4126 found++;
4127 }
4128 if (found)
4129 return;
4130
4131 /* otherwise we will clear the whole queue
4132 * by copying the empty outq at the end of this
4133 * function, but first clear the current entries
4134 * in the queue.
4135 */
4136 for (i = 0; i < q->nr; i++)
4137 diff_free_filepair(q->queue[i]);
4138 }
4139 else {
4140 /* Only the matching ones */
4141 for (i = 0; i < q->nr; i++) {
4142 struct diff_filepair *p = q->queue[i];
4143
4144 if (((p->status == DIFF_STATUS_MODIFIED) &&
4145 ((p->score &&
4146 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4147 (!p->score &&
4148 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4149 ((p->status != DIFF_STATUS_MODIFIED) &&
4150 strchr(filter, p->status)))
4151 diff_q(&outq, p);
4152 else
4153 diff_free_filepair(p);
4154 }
4155 }
4156 free(q->queue);
4157 *q = outq;
4158 }
4159
4160 /* Check whether two filespecs with the same mode and size are identical */
4161 static int diff_filespec_is_identical(struct diff_filespec *one,
4162 struct diff_filespec *two)
4163 {
4164 if (S_ISGITLINK(one->mode))
4165 return 0;
4166 if (diff_populate_filespec(one, 0))
4167 return 0;
4168 if (diff_populate_filespec(two, 0))
4169 return 0;
4170 return !memcmp(one->data, two->data, one->size);
4171 }
4172
4173 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4174 {
4175 int i;
4176 struct diff_queue_struct *q = &diff_queued_diff;
4177 struct diff_queue_struct outq;
4178 DIFF_QUEUE_CLEAR(&outq);
4179
4180 for (i = 0; i < q->nr; i++) {
4181 struct diff_filepair *p = q->queue[i];
4182
4183 /*
4184 * 1. Entries that come from stat info dirtiness
4185 * always have both sides (iow, not create/delete),
4186 * one side of the object name is unknown, with
4187 * the same mode and size. Keep the ones that
4188 * do not match these criteria. They have real
4189 * differences.
4190 *
4191 * 2. At this point, the file is known to be modified,
4192 * with the same mode and size, and the object
4193 * name of one side is unknown. Need to inspect
4194 * the identical contents.
4195 */
4196 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4197 !DIFF_FILE_VALID(p->two) ||
4198 (p->one->sha1_valid && p->two->sha1_valid) ||
4199 (p->one->mode != p->two->mode) ||
4200 diff_populate_filespec(p->one, 1) ||
4201 diff_populate_filespec(p->two, 1) ||
4202 (p->one->size != p->two->size) ||
4203 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4204 diff_q(&outq, p);
4205 else {
4206 /*
4207 * The caller can subtract 1 from skip_stat_unmatch
4208 * to determine how many paths were dirty only
4209 * due to stat info mismatch.
4210 */
4211 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4212 diffopt->skip_stat_unmatch++;
4213 diff_free_filepair(p);
4214 }
4215 }
4216 free(q->queue);
4217 *q = outq;
4218 }
4219
4220 static int diffnamecmp(const void *a_, const void *b_)
4221 {
4222 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4223 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4224 const char *name_a, *name_b;
4225
4226 name_a = a->one ? a->one->path : a->two->path;
4227 name_b = b->one ? b->one->path : b->two->path;
4228 return strcmp(name_a, name_b);
4229 }
4230
4231 void diffcore_fix_diff_index(struct diff_options *options)
4232 {
4233 struct diff_queue_struct *q = &diff_queued_diff;
4234 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4235 }
4236
4237 void diffcore_std(struct diff_options *options)
4238 {
4239 if (options->skip_stat_unmatch)
4240 diffcore_skip_stat_unmatch(options);
4241 if (!options->found_follow) {
4242 /* See try_to_follow_renames() in tree-diff.c */
4243 if (options->break_opt != -1)
4244 diffcore_break(options->break_opt);
4245 if (options->detect_rename)
4246 diffcore_rename(options);
4247 if (options->break_opt != -1)
4248 diffcore_merge_broken();
4249 }
4250 if (options->pickaxe)
4251 diffcore_pickaxe(options);
4252 if (options->orderfile)
4253 diffcore_order(options->orderfile);
4254 if (!options->found_follow)
4255 /* See try_to_follow_renames() in tree-diff.c */
4256 diff_resolve_rename_copy();
4257 diffcore_apply_filter(options->filter);
4258
4259 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4260 DIFF_OPT_SET(options, HAS_CHANGES);
4261 else
4262 DIFF_OPT_CLR(options, HAS_CHANGES);
4263
4264 options->found_follow = 0;
4265 }
4266
4267 int diff_result_code(struct diff_options *opt, int status)
4268 {
4269 int result = 0;
4270
4271 diff_warn_rename_limit("diff.renamelimit",
4272 opt->needed_rename_limit,
4273 opt->degraded_cc_to_c);
4274 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4275 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4276 return status;
4277 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4278 DIFF_OPT_TST(opt, HAS_CHANGES))
4279 result |= 01;
4280 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4281 DIFF_OPT_TST(opt, CHECK_FAILED))
4282 result |= 02;
4283 return result;
4284 }
4285
4286 /*
4287 * Shall changes to this submodule be ignored?
4288 *
4289 * Submodule changes can be configured to be ignored separately for each path,
4290 * but that configuration can be overridden from the command line.
4291 */
4292 static int is_submodule_ignored(const char *path, struct diff_options *options)
4293 {
4294 int ignored = 0;
4295 unsigned orig_flags = options->flags;
4296 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4297 set_diffopt_flags_from_submodule_config(options, path);
4298 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4299 ignored = 1;
4300 options->flags = orig_flags;
4301 return ignored;
4302 }
4303
4304 void diff_addremove(struct diff_options *options,
4305 int addremove, unsigned mode,
4306 const unsigned char *sha1,
4307 const char *concatpath, unsigned dirty_submodule)
4308 {
4309 struct diff_filespec *one, *two;
4310
4311 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4312 return;
4313
4314 /* This may look odd, but it is a preparation for
4315 * feeding "there are unchanged files which should
4316 * not produce diffs, but when you are doing copy
4317 * detection you would need them, so here they are"
4318 * entries to the diff-core. They will be prefixed
4319 * with something like '=' or '*' (I haven't decided
4320 * which but should not make any difference).
4321 * Feeding the same new and old to diff_change()
4322 * also has the same effect.
4323 * Before the final output happens, they are pruned after
4324 * merged into rename/copy pairs as appropriate.
4325 */
4326 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4327 addremove = (addremove == '+' ? '-' :
4328 addremove == '-' ? '+' : addremove);
4329
4330 if (options->prefix &&
4331 strncmp(concatpath, options->prefix, options->prefix_length))
4332 return;
4333
4334 one = alloc_filespec(concatpath);
4335 two = alloc_filespec(concatpath);
4336
4337 if (addremove != '+')
4338 fill_filespec(one, sha1, mode);
4339 if (addremove != '-') {
4340 fill_filespec(two, sha1, mode);
4341 two->dirty_submodule = dirty_submodule;
4342 }
4343
4344 diff_queue(&diff_queued_diff, one, two);
4345 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4346 DIFF_OPT_SET(options, HAS_CHANGES);
4347 }
4348
4349 void diff_change(struct diff_options *options,
4350 unsigned old_mode, unsigned new_mode,
4351 const unsigned char *old_sha1,
4352 const unsigned char *new_sha1,
4353 const char *concatpath,
4354 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4355 {
4356 struct diff_filespec *one, *two;
4357
4358 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4359 is_submodule_ignored(concatpath, options))
4360 return;
4361
4362 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4363 unsigned tmp;
4364 const unsigned char *tmp_c;
4365 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4366 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4367 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4368 new_dirty_submodule = tmp;
4369 }
4370
4371 if (options->prefix &&
4372 strncmp(concatpath, options->prefix, options->prefix_length))
4373 return;
4374
4375 one = alloc_filespec(concatpath);
4376 two = alloc_filespec(concatpath);
4377 fill_filespec(one, old_sha1, old_mode);
4378 fill_filespec(two, new_sha1, new_mode);
4379 one->dirty_submodule = old_dirty_submodule;
4380 two->dirty_submodule = new_dirty_submodule;
4381
4382 diff_queue(&diff_queued_diff, one, two);
4383 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4384 DIFF_OPT_SET(options, HAS_CHANGES);
4385 }
4386
4387 void diff_unmerge(struct diff_options *options,
4388 const char *path,
4389 unsigned mode, const unsigned char *sha1)
4390 {
4391 struct diff_filespec *one, *two;
4392
4393 if (options->prefix &&
4394 strncmp(path, options->prefix, options->prefix_length))
4395 return;
4396
4397 one = alloc_filespec(path);
4398 two = alloc_filespec(path);
4399 fill_filespec(one, sha1, mode);
4400 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4401 }
4402
4403 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4404 size_t *outsize)
4405 {
4406 struct diff_tempfile *temp;
4407 const char *argv[3];
4408 const char **arg = argv;
4409 struct child_process child;
4410 struct strbuf buf = STRBUF_INIT;
4411 int err = 0;
4412
4413 temp = prepare_temp_file(spec->path, spec);
4414 *arg++ = pgm;
4415 *arg++ = temp->name;
4416 *arg = NULL;
4417
4418 memset(&child, 0, sizeof(child));
4419 child.use_shell = 1;
4420 child.argv = argv;
4421 child.out = -1;
4422 if (start_command(&child)) {
4423 remove_tempfile();
4424 return NULL;
4425 }
4426
4427 if (strbuf_read(&buf, child.out, 0) < 0)
4428 err = error("error reading from textconv command '%s'", pgm);
4429 close(child.out);
4430
4431 if (finish_command(&child) || err) {
4432 strbuf_release(&buf);
4433 remove_tempfile();
4434 return NULL;
4435 }
4436 remove_tempfile();
4437
4438 return strbuf_detach(&buf, outsize);
4439 }
4440
4441 size_t fill_textconv(struct userdiff_driver *driver,
4442 struct diff_filespec *df,
4443 char **outbuf)
4444 {
4445 size_t size;
4446
4447 if (!driver || !driver->textconv) {
4448 if (!DIFF_FILE_VALID(df)) {
4449 *outbuf = "";
4450 return 0;
4451 }
4452 if (diff_populate_filespec(df, 0))
4453 die("unable to read files to diff");
4454 *outbuf = df->data;
4455 return df->size;
4456 }
4457
4458 if (driver->textconv_cache && df->sha1_valid) {
4459 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4460 &size);
4461 if (*outbuf)
4462 return size;
4463 }
4464
4465 *outbuf = run_textconv(driver->textconv, df, &size);
4466 if (!*outbuf)
4467 die("unable to read files to diff");
4468
4469 if (driver->textconv_cache && df->sha1_valid) {
4470 /* ignore errors, as we might be in a readonly repository */
4471 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4472 size);
4473 /*
4474 * we could save up changes and flush them all at the end,
4475 * but we would need an extra call after all diffing is done.
4476 * Since generating a cache entry is the slow path anyway,
4477 * this extra overhead probably isn't a big deal.
4478 */
4479 notes_cache_write(driver->textconv_cache);
4480 }
4481
4482 return size;
4483 }
Dscho's miscellaneous git branches