?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff: introduce --stat-lines to limit the stat lines
[git.git] / diff.c
blob82789e30eb7091d9405635438b6c406aac4adb40
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, count;
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 count = options->stat_count ? options->stat_count : data->nr;
1263
1264 /* Sanity: give at least 5 columns to the graph,
1265 * but leave at least 10 columns for the name.
1266 */
1267 if (width < 25)
1268 width = 25;
1269 if (name_width < 10)
1270 name_width = 10;
1271 else if (width < name_width + 15)
1272 name_width = width - 15;
1273
1274 /* Find the longest filename and max number of changes */
1275 reset = diff_get_color_opt(options, DIFF_RESET);
1276 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1277 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1278
1279 for (i = 0; (i < count) && (i < data->nr); i++) {
1280 struct diffstat_file *file = data->files[i];
1281 uintmax_t change = file->added + file->deleted;
1282 if (!data->files[i]->is_renamed &&
1283 (change == 0)) {
1284 count++; /* not shown == room for one more */
1285 continue;
1286 }
1287 fill_print_name(file);
1288 len = strlen(file->print_name);
1289 if (max_len < len)
1290 max_len = len;
1291
1292 if (file->is_binary || file->is_unmerged)
1293 continue;
1294 if (max_change < change)
1295 max_change = change;
1296 }
1297 count = i; /* min(count, data->nr) */
1298
1299 /* Compute the width of the graph part;
1300 * 10 is for one blank at the beginning of the line plus
1301 * " | count " between the name and the graph.
1302 *
1303 * From here on, name_width is the width of the name area,
1304 * and width is the width of the graph area.
1305 */
1306 name_width = (name_width < max_len) ? name_width : max_len;
1307 if (width < (name_width + 10) + max_change)
1308 width = width - (name_width + 10);
1309 else
1310 width = max_change;
1311
1312 for (i = 0; i < count; i++) {
1313 const char *prefix = "";
1314 char *name = data->files[i]->print_name;
1315 uintmax_t added = data->files[i]->added;
1316 uintmax_t deleted = data->files[i]->deleted;
1317 int name_len;
1318
1319 if (!data->files[i]->is_renamed &&
1320 (added + deleted == 0)) {
1321 total_files--;
1322 continue;
1323 }
1324 /*
1325 * "scale" the filename
1326 */
1327 len = name_width;
1328 name_len = strlen(name);
1329 if (name_width < name_len) {
1330 char *slash;
1331 prefix = "...";
1332 len -= 3;
1333 name += name_len - len;
1334 slash = strchr(name, '/');
1335 if (slash)
1336 name = slash;
1337 }
1338
1339 if (data->files[i]->is_binary) {
1340 fprintf(options->file, "%s", line_prefix);
1341 show_name(options->file, prefix, name, len);
1342 fprintf(options->file, " Bin ");
1343 fprintf(options->file, "%s%"PRIuMAX"%s",
1344 del_c, deleted, reset);
1345 fprintf(options->file, " -> ");
1346 fprintf(options->file, "%s%"PRIuMAX"%s",
1347 add_c, added, reset);
1348 fprintf(options->file, " bytes");
1349 fprintf(options->file, "\n");
1350 continue;
1351 }
1352 else if (data->files[i]->is_unmerged) {
1353 fprintf(options->file, "%s", line_prefix);
1354 show_name(options->file, prefix, name, len);
1355 fprintf(options->file, " Unmerged\n");
1356 continue;
1357 }
1358
1359 /*
1360 * scale the add/delete
1361 */
1362 add = added;
1363 del = deleted;
1364 adds += add;
1365 dels += del;
1366
1367 if (width <= max_change) {
1368 add = scale_linear(add, width, max_change);
1369 del = scale_linear(del, width, max_change);
1370 }
1371 fprintf(options->file, "%s", line_prefix);
1372 show_name(options->file, prefix, name, len);
1373 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1374 added + deleted ? " " : "");
1375 show_graph(options->file, '+', add, add_c, reset);
1376 show_graph(options->file, '-', del, del_c, reset);
1377 fprintf(options->file, "\n");
1378 }
1379 if (count < data->nr)
1380 fprintf(options->file, "%s ...\n", line_prefix);
1381 for (i = count; i < data->nr; i++) {
1382 uintmax_t added = data->files[i]->added;
1383 uintmax_t deleted = data->files[i]->deleted;
1384 if (!data->files[i]->is_renamed &&
1385 (added + deleted == 0)) {
1386 total_files--;
1387 continue;
1388 }
1389 adds += added;
1390 dels += deleted;
1391 }
1392 fprintf(options->file, "%s", line_prefix);
1393 fprintf(options->file,
1394 " %d files changed, %d insertions(+), %d deletions(-)\n",
1395 total_files, adds, dels);
1396 }
1397
1398 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1399 {
1400 int i, adds = 0, dels = 0, total_files = data->nr;
1401
1402 if (data->nr == 0)
1403 return;
1404
1405 for (i = 0; i < data->nr; i++) {
1406 if (!data->files[i]->is_binary &&
1407 !data->files[i]->is_unmerged) {
1408 int added = data->files[i]->added;
1409 int deleted= data->files[i]->deleted;
1410 if (!data->files[i]->is_renamed &&
1411 (added + deleted == 0)) {
1412 total_files--;
1413 } else {
1414 adds += added;
1415 dels += deleted;
1416 }
1417 }
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 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1426 total_files, adds, dels);
1427 }
1428
1429 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1430 {
1431 int i;
1432
1433 if (data->nr == 0)
1434 return;
1435
1436 for (i = 0; i < data->nr; i++) {
1437 struct diffstat_file *file = data->files[i];
1438
1439 if (options->output_prefix) {
1440 struct strbuf *msg = NULL;
1441 msg = options->output_prefix(options,
1442 options->output_prefix_data);
1443 fprintf(options->file, "%s", msg->buf);
1444 }
1445
1446 if (file->is_binary)
1447 fprintf(options->file, "-\t-\t");
1448 else
1449 fprintf(options->file,
1450 "%"PRIuMAX"\t%"PRIuMAX"\t",
1451 file->added, file->deleted);
1452 if (options->line_termination) {
1453 fill_print_name(file);
1454 if (!file->is_renamed)
1455 write_name_quoted(file->name, options->file,
1456 options->line_termination);
1457 else {
1458 fputs(file->print_name, options->file);
1459 putc(options->line_termination, options->file);
1460 }
1461 } else {
1462 if (file->is_renamed) {
1463 putc('\0', options->file);
1464 write_name_quoted(file->from_name, options->file, '\0');
1465 }
1466 write_name_quoted(file->name, options->file, '\0');
1467 }
1468 }
1469 }
1470
1471 struct dirstat_file {
1472 const char *name;
1473 unsigned long changed;
1474 };
1475
1476 struct dirstat_dir {
1477 struct dirstat_file *files;
1478 int alloc, nr, percent, cumulative;
1479 };
1480
1481 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1482 unsigned long changed, const char *base, int baselen)
1483 {
1484 unsigned long this_dir = 0;
1485 unsigned int sources = 0;
1486 const char *line_prefix = "";
1487 struct strbuf *msg = NULL;
1488
1489 if (opt->output_prefix) {
1490 msg = opt->output_prefix(opt, opt->output_prefix_data);
1491 line_prefix = msg->buf;
1492 }
1493
1494 while (dir->nr) {
1495 struct dirstat_file *f = dir->files;
1496 int namelen = strlen(f->name);
1497 unsigned long this;
1498 char *slash;
1499
1500 if (namelen < baselen)
1501 break;
1502 if (memcmp(f->name, base, baselen))
1503 break;
1504 slash = strchr(f->name + baselen, '/');
1505 if (slash) {
1506 int newbaselen = slash + 1 - f->name;
1507 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1508 sources++;
1509 } else {
1510 this = f->changed;
1511 dir->files++;
1512 dir->nr--;
1513 sources += 2;
1514 }
1515 this_dir += this;
1516 }
1517
1518 /*
1519 * We don't report dirstat's for
1520 * - the top level
1521 * - or cases where everything came from a single directory
1522 * under this directory (sources == 1).
1523 */
1524 if (baselen && sources != 1) {
1525 int permille = this_dir * 1000 / changed;
1526 if (permille) {
1527 int percent = permille / 10;
1528 if (percent >= dir->percent) {
1529 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1530 percent, permille % 10, baselen, base);
1531 if (!dir->cumulative)
1532 return 0;
1533 }
1534 }
1535 }
1536 return this_dir;
1537 }
1538
1539 static int dirstat_compare(const void *_a, const void *_b)
1540 {
1541 const struct dirstat_file *a = _a;
1542 const struct dirstat_file *b = _b;
1543 return strcmp(a->name, b->name);
1544 }
1545
1546 static void show_dirstat(struct diff_options *options)
1547 {
1548 int i;
1549 unsigned long changed;
1550 struct dirstat_dir dir;
1551 struct diff_queue_struct *q = &diff_queued_diff;
1552
1553 dir.files = NULL;
1554 dir.alloc = 0;
1555 dir.nr = 0;
1556 dir.percent = options->dirstat_percent;
1557 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1558
1559 changed = 0;
1560 for (i = 0; i < q->nr; i++) {
1561 struct diff_filepair *p = q->queue[i];
1562 const char *name;
1563 unsigned long copied, added, damage;
1564 int content_changed;
1565
1566 name = p->two->path ? p->two->path : p->one->path;
1567
1568 if (p->one->sha1_valid && p->two->sha1_valid)
1569 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1570 else
1571 content_changed = 1;
1572
1573 if (!content_changed) {
1574 /*
1575 * The SHA1 has not changed, so pre-/post-content is
1576 * identical. We can therefore skip looking at the
1577 * file contents altogether.
1578 */
1579 damage = 0;
1580 goto found_damage;
1581 }
1582
1583 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1584 /*
1585 * In --dirstat-by-file mode, we don't really need to
1586 * look at the actual file contents at all.
1587 * The fact that the SHA1 changed is enough for us to
1588 * add this file to the list of results
1589 * (with each file contributing equal damage).
1590 */
1591 damage = 1;
1592 goto found_damage;
1593 }
1594
1595 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1596 diff_populate_filespec(p->one, 0);
1597 diff_populate_filespec(p->two, 0);
1598 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1599 &copied, &added);
1600 diff_free_filespec_data(p->one);
1601 diff_free_filespec_data(p->two);
1602 } else if (DIFF_FILE_VALID(p->one)) {
1603 diff_populate_filespec(p->one, 1);
1604 copied = added = 0;
1605 diff_free_filespec_data(p->one);
1606 } else if (DIFF_FILE_VALID(p->two)) {
1607 diff_populate_filespec(p->two, 1);
1608 copied = 0;
1609 added = p->two->size;
1610 diff_free_filespec_data(p->two);
1611 } else
1612 continue;
1613
1614 /*
1615 * Original minus copied is the removed material,
1616 * added is the new material. They are both damages
1617 * made to the preimage.
1618 * If the resulting damage is zero, we know that
1619 * diffcore_count_changes() considers the two entries to
1620 * be identical, but since content_changed is true, we
1621 * know that there must have been _some_ kind of change,
1622 * so we force all entries to have damage > 0.
1623 */
1624 damage = (p->one->size - copied) + added;
1625 if (!damage)
1626 damage = 1;
1627
1628 found_damage:
1629 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1630 dir.files[dir.nr].name = name;
1631 dir.files[dir.nr].changed = damage;
1632 changed += damage;
1633 dir.nr++;
1634 }
1635
1636 /* This can happen even with many files, if everything was renames */
1637 if (!changed)
1638 return;
1639
1640 /* Show all directories with more than x% of the changes */
1641 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1642 gather_dirstat(options, &dir, changed, "", 0);
1643 }
1644
1645 static void free_diffstat_info(struct diffstat_t *diffstat)
1646 {
1647 int i;
1648 for (i = 0; i < diffstat->nr; i++) {
1649 struct diffstat_file *f = diffstat->files[i];
1650 if (f->name != f->print_name)
1651 free(f->print_name);
1652 free(f->name);
1653 free(f->from_name);
1654 free(f);
1655 }
1656 free(diffstat->files);
1657 }
1658
1659 struct checkdiff_t {
1660 const char *filename;
1661 int lineno;
1662 int conflict_marker_size;
1663 struct diff_options *o;
1664 unsigned ws_rule;
1665 unsigned status;
1666 };
1667
1668 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1669 {
1670 char firstchar;
1671 int cnt;
1672
1673 if (len < marker_size + 1)
1674 return 0;
1675 firstchar = line[0];
1676 switch (firstchar) {
1677 case '=': case '>': case '<': case '|':
1678 break;
1679 default:
1680 return 0;
1681 }
1682 for (cnt = 1; cnt < marker_size; cnt++)
1683 if (line[cnt] != firstchar)
1684 return 0;
1685 /* line[1] thru line[marker_size-1] are same as firstchar */
1686 if (len < marker_size + 1 || !isspace(line[marker_size]))
1687 return 0;
1688 return 1;
1689 }
1690
1691 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1692 {
1693 struct checkdiff_t *data = priv;
1694 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1695 int marker_size = data->conflict_marker_size;
1696 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1697 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1698 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1699 char *err;
1700 char *line_prefix = "";
1701 struct strbuf *msgbuf;
1702
1703 assert(data->o);
1704 if (data->o->output_prefix) {
1705 msgbuf = data->o->output_prefix(data->o,
1706 data->o->output_prefix_data);
1707 line_prefix = msgbuf->buf;
1708 }
1709
1710 if (line[0] == '+') {
1711 unsigned bad;
1712 data->lineno++;
1713 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1714 data->status |= 1;
1715 fprintf(data->o->file,
1716 "%s%s:%d: leftover conflict marker\n",
1717 line_prefix, data->filename, data->lineno);
1718 }
1719 bad = ws_check(line + 1, len - 1, data->ws_rule);
1720 if (!bad)
1721 return;
1722 data->status |= bad;
1723 err = whitespace_error_string(bad);
1724 fprintf(data->o->file, "%s%s:%d: %s.\n",
1725 line_prefix, data->filename, data->lineno, err);
1726 free(err);
1727 emit_line(data->o, set, reset, line, 1);
1728 ws_check_emit(line + 1, len - 1, data->ws_rule,
1729 data->o->file, set, reset, ws);
1730 } else if (line[0] == ' ') {
1731 data->lineno++;
1732 } else if (line[0] == '@') {
1733 char *plus = strchr(line, '+');
1734 if (plus)
1735 data->lineno = strtol(plus, NULL, 10) - 1;
1736 else
1737 die("invalid diff");
1738 }
1739 }
1740
1741 static unsigned char *deflate_it(char *data,
1742 unsigned long size,
1743 unsigned long *result_size)
1744 {
1745 int bound;
1746 unsigned char *deflated;
1747 z_stream stream;
1748
1749 memset(&stream, 0, sizeof(stream));
1750 deflateInit(&stream, zlib_compression_level);
1751 bound = deflateBound(&stream, size);
1752 deflated = xmalloc(bound);
1753 stream.next_out = deflated;
1754 stream.avail_out = bound;
1755
1756 stream.next_in = (unsigned char *)data;
1757 stream.avail_in = size;
1758 while (deflate(&stream, Z_FINISH) == Z_OK)
1759 ; /* nothing */
1760 deflateEnd(&stream);
1761 *result_size = stream.total_out;
1762 return deflated;
1763 }
1764
1765 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1766 {
1767 void *cp;
1768 void *delta;
1769 void *deflated;
1770 void *data;
1771 unsigned long orig_size;
1772 unsigned long delta_size;
1773 unsigned long deflate_size;
1774 unsigned long data_size;
1775
1776 /* We could do deflated delta, or we could do just deflated two,
1777 * whichever is smaller.
1778 */
1779 delta = NULL;
1780 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1781 if (one->size && two->size) {
1782 delta = diff_delta(one->ptr, one->size,
1783 two->ptr, two->size,
1784 &delta_size, deflate_size);
1785 if (delta) {
1786 void *to_free = delta;
1787 orig_size = delta_size;
1788 delta = deflate_it(delta, delta_size, &delta_size);
1789 free(to_free);
1790 }
1791 }
1792
1793 if (delta && delta_size < deflate_size) {
1794 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1795 free(deflated);
1796 data = delta;
1797 data_size = delta_size;
1798 }
1799 else {
1800 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1801 free(delta);
1802 data = deflated;
1803 data_size = deflate_size;
1804 }
1805
1806 /* emit data encoded in base85 */
1807 cp = data;
1808 while (data_size) {
1809 int bytes = (52 < data_size) ? 52 : data_size;
1810 char line[70];
1811 data_size -= bytes;
1812 if (bytes <= 26)
1813 line[0] = bytes + 'A' - 1;
1814 else
1815 line[0] = bytes - 26 + 'a' - 1;
1816 encode_85(line + 1, cp, bytes);
1817 cp = (char *) cp + bytes;
1818 fprintf(file, "%s", prefix);
1819 fputs(line, file);
1820 fputc('\n', file);
1821 }
1822 fprintf(file, "%s\n", prefix);
1823 free(data);
1824 }
1825
1826 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1827 {
1828 fprintf(file, "%sGIT binary patch\n", prefix);
1829 emit_binary_diff_body(file, one, two, prefix);
1830 emit_binary_diff_body(file, two, one, prefix);
1831 }
1832
1833 static void diff_filespec_load_driver(struct diff_filespec *one)
1834 {
1835 /* Use already-loaded driver */
1836 if (one->driver)
1837 return;
1838
1839 if (S_ISREG(one->mode))
1840 one->driver = userdiff_find_by_path(one->path);
1841
1842 /* Fallback to default settings */
1843 if (!one->driver)
1844 one->driver = userdiff_find_by_name("default");
1845 }
1846
1847 int diff_filespec_is_binary(struct diff_filespec *one)
1848 {
1849 if (one->is_binary == -1) {
1850 diff_filespec_load_driver(one);
1851 if (one->driver->binary != -1)
1852 one->is_binary = one->driver->binary;
1853 else {
1854 if (!one->data && DIFF_FILE_VALID(one))
1855 diff_populate_filespec(one, 0);
1856 if (one->data)
1857 one->is_binary = buffer_is_binary(one->data,
1858 one->size);
1859 if (one->is_binary == -1)
1860 one->is_binary = 0;
1861 }
1862 }
1863 return one->is_binary;
1864 }
1865
1866 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1867 {
1868 diff_filespec_load_driver(one);
1869 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1870 }
1871
1872 static const char *userdiff_word_regex(struct diff_filespec *one)
1873 {
1874 diff_filespec_load_driver(one);
1875 return one->driver->word_regex;
1876 }
1877
1878 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1879 {
1880 if (!options->a_prefix)
1881 options->a_prefix = a;
1882 if (!options->b_prefix)
1883 options->b_prefix = b;
1884 }
1885
1886 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1887 {
1888 if (!DIFF_FILE_VALID(one))
1889 return NULL;
1890
1891 diff_filespec_load_driver(one);
1892 if (!one->driver->textconv)
1893 return NULL;
1894
1895 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1896 struct notes_cache *c = xmalloc(sizeof(*c));
1897 struct strbuf name = STRBUF_INIT;
1898
1899 strbuf_addf(&name, "textconv/%s", one->driver->name);
1900 notes_cache_init(c, name.buf, one->driver->textconv);
1901 one->driver->textconv_cache = c;
1902 }
1903
1904 return one->driver;
1905 }
1906
1907 static void builtin_diff(const char *name_a,
1908 const char *name_b,
1909 struct diff_filespec *one,
1910 struct diff_filespec *two,
1911 const char *xfrm_msg,
1912 int must_show_header,
1913 struct diff_options *o,
1914 int complete_rewrite)
1915 {
1916 mmfile_t mf1, mf2;
1917 const char *lbl[2];
1918 char *a_one, *b_two;
1919 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1920 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1921 const char *a_prefix, *b_prefix;
1922 struct userdiff_driver *textconv_one = NULL;
1923 struct userdiff_driver *textconv_two = NULL;
1924 struct strbuf header = STRBUF_INIT;
1925 struct strbuf *msgbuf;
1926 char *line_prefix = "";
1927
1928 if (o->output_prefix) {
1929 msgbuf = o->output_prefix(o, o->output_prefix_data);
1930 line_prefix = msgbuf->buf;
1931 }
1932
1933 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1934 (!one->mode || S_ISGITLINK(one->mode)) &&
1935 (!two->mode || S_ISGITLINK(two->mode))) {
1936 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1937 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1938 show_submodule_summary(o->file, one ? one->path : two->path,
1939 one->sha1, two->sha1, two->dirty_submodule,
1940 del, add, reset);
1941 return;
1942 }
1943
1944 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1945 textconv_one = get_textconv(one);
1946 textconv_two = get_textconv(two);
1947 }
1948
1949 diff_set_mnemonic_prefix(o, "a/", "b/");
1950 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1951 a_prefix = o->b_prefix;
1952 b_prefix = o->a_prefix;
1953 } else {
1954 a_prefix = o->a_prefix;
1955 b_prefix = o->b_prefix;
1956 }
1957
1958 /* Never use a non-valid filename anywhere if at all possible */
1959 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1960 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1961
1962 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1963 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1964 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1965 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1966 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1967 if (lbl[0][0] == '/') {
1968 /* /dev/null */
1969 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1970 if (xfrm_msg)
1971 strbuf_addstr(&header, xfrm_msg);
1972 must_show_header = 1;
1973 }
1974 else if (lbl[1][0] == '/') {
1975 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1976 if (xfrm_msg)
1977 strbuf_addstr(&header, xfrm_msg);
1978 must_show_header = 1;
1979 }
1980 else {
1981 if (one->mode != two->mode) {
1982 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1983 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1984 must_show_header = 1;
1985 }
1986 if (xfrm_msg)
1987 strbuf_addstr(&header, xfrm_msg);
1988
1989 /*
1990 * we do not run diff between different kind
1991 * of objects.
1992 */
1993 if ((one->mode ^ two->mode) & S_IFMT)
1994 goto free_ab_and_return;
1995 if (complete_rewrite &&
1996 (textconv_one || !diff_filespec_is_binary(one)) &&
1997 (textconv_two || !diff_filespec_is_binary(two))) {
1998 fprintf(o->file, "%s", header.buf);
1999 strbuf_reset(&header);
2000 emit_rewrite_diff(name_a, name_b, one, two,
2001 textconv_one, textconv_two, o);
2002 o->found_changes = 1;
2003 goto free_ab_and_return;
2004 }
2005 }
2006
2007 if (o->irreversible_delete && lbl[1][0] == '/') {
2008 fprintf(o->file, "%s", header.buf);
2009 strbuf_reset(&header);
2010 goto free_ab_and_return;
2011 } else if (!DIFF_OPT_TST(o, TEXT) &&
2012 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2013 (!textconv_two && diff_filespec_is_binary(two)) )) {
2014 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2015 die("unable to read files to diff");
2016 /* Quite common confusing case */
2017 if (mf1.size == mf2.size &&
2018 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2019 if (must_show_header)
2020 fprintf(o->file, "%s", header.buf);
2021 goto free_ab_and_return;
2022 }
2023 fprintf(o->file, "%s", header.buf);
2024 strbuf_reset(&header);
2025 if (DIFF_OPT_TST(o, BINARY))
2026 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2027 else
2028 fprintf(o->file, "%sBinary files %s and %s differ\n",
2029 line_prefix, lbl[0], lbl[1]);
2030 o->found_changes = 1;
2031 } else {
2032 /* Crazy xdl interfaces.. */
2033 const char *diffopts = getenv("GIT_DIFF_OPTS");
2034 xpparam_t xpp;
2035 xdemitconf_t xecfg;
2036 struct emit_callback ecbdata;
2037 const struct userdiff_funcname *pe;
2038
2039 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2040 fprintf(o->file, "%s", header.buf);
2041 strbuf_reset(&header);
2042 }
2043
2044 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2045 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2046
2047 pe = diff_funcname_pattern(one);
2048 if (!pe)
2049 pe = diff_funcname_pattern(two);
2050
2051 memset(&xpp, 0, sizeof(xpp));
2052 memset(&xecfg, 0, sizeof(xecfg));
2053 memset(&ecbdata, 0, sizeof(ecbdata));
2054 ecbdata.label_path = lbl;
2055 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2056 ecbdata.found_changesp = &o->found_changes;
2057 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2058 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2059 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2060 ecbdata.opt = o;
2061 ecbdata.header = header.len ? &header : NULL;
2062 xpp.flags = o->xdl_opts;
2063 xecfg.ctxlen = o->context;
2064 xecfg.interhunkctxlen = o->interhunkcontext;
2065 xecfg.flags = XDL_EMIT_FUNCNAMES;
2066 if (pe)
2067 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2068 if (!diffopts)
2069 ;
2070 else if (!prefixcmp(diffopts, "--unified="))
2071 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2072 else if (!prefixcmp(diffopts, "-u"))
2073 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2074 if (o->word_diff) {
2075 int i;
2076
2077 ecbdata.diff_words =
2078 xcalloc(1, sizeof(struct diff_words_data));
2079 ecbdata.diff_words->type = o->word_diff;
2080 ecbdata.diff_words->opt = o;
2081 if (!o->word_regex)
2082 o->word_regex = userdiff_word_regex(one);
2083 if (!o->word_regex)
2084 o->word_regex = userdiff_word_regex(two);
2085 if (!o->word_regex)
2086 o->word_regex = diff_word_regex_cfg;
2087 if (o->word_regex) {
2088 ecbdata.diff_words->word_regex = (regex_t *)
2089 xmalloc(sizeof(regex_t));
2090 if (regcomp(ecbdata.diff_words->word_regex,
2091 o->word_regex,
2092 REG_EXTENDED | REG_NEWLINE))
2093 die ("Invalid regular expression: %s",
2094 o->word_regex);
2095 }
2096 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2097 if (o->word_diff == diff_words_styles[i].type) {
2098 ecbdata.diff_words->style =
2099 &diff_words_styles[i];
2100 break;
2101 }
2102 }
2103 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2104 struct diff_words_style *st = ecbdata.diff_words->style;
2105 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2106 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2107 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2108 }
2109 }
2110 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2111 &xpp, &xecfg);
2112 if (o->word_diff)
2113 free_diff_words_data(&ecbdata);
2114 if (textconv_one)
2115 free(mf1.ptr);
2116 if (textconv_two)
2117 free(mf2.ptr);
2118 xdiff_clear_find_func(&xecfg);
2119 }
2120
2121 free_ab_and_return:
2122 strbuf_release(&header);
2123 diff_free_filespec_data(one);
2124 diff_free_filespec_data(two);
2125 free(a_one);
2126 free(b_two);
2127 return;
2128 }
2129
2130 static void builtin_diffstat(const char *name_a, const char *name_b,
2131 struct diff_filespec *one,
2132 struct diff_filespec *two,
2133 struct diffstat_t *diffstat,
2134 struct diff_options *o,
2135 int complete_rewrite)
2136 {
2137 mmfile_t mf1, mf2;
2138 struct diffstat_file *data;
2139
2140 data = diffstat_add(diffstat, name_a, name_b);
2141
2142 if (!one || !two) {
2143 data->is_unmerged = 1;
2144 return;
2145 }
2146
2147 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2148 data->is_binary = 1;
2149 data->added = diff_filespec_size(two);
2150 data->deleted = diff_filespec_size(one);
2151 }
2152
2153 else if (complete_rewrite) {
2154 diff_populate_filespec(one, 0);
2155 diff_populate_filespec(two, 0);
2156 data->deleted = count_lines(one->data, one->size);
2157 data->added = count_lines(two->data, two->size);
2158 }
2159
2160 else {
2161 /* Crazy xdl interfaces.. */
2162 xpparam_t xpp;
2163 xdemitconf_t xecfg;
2164
2165 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2166 die("unable to read files to diff");
2167
2168 memset(&xpp, 0, sizeof(xpp));
2169 memset(&xecfg, 0, sizeof(xecfg));
2170 xpp.flags = o->xdl_opts;
2171 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2172 &xpp, &xecfg);
2173 }
2174
2175 diff_free_filespec_data(one);
2176 diff_free_filespec_data(two);
2177 }
2178
2179 static void builtin_checkdiff(const char *name_a, const char *name_b,
2180 const char *attr_path,
2181 struct diff_filespec *one,
2182 struct diff_filespec *two,
2183 struct diff_options *o)
2184 {
2185 mmfile_t mf1, mf2;
2186 struct checkdiff_t data;
2187
2188 if (!two)
2189 return;
2190
2191 memset(&data, 0, sizeof(data));
2192 data.filename = name_b ? name_b : name_a;
2193 data.lineno = 0;
2194 data.o = o;
2195 data.ws_rule = whitespace_rule(attr_path);
2196 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2197
2198 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2199 die("unable to read files to diff");
2200
2201 /*
2202 * All the other codepaths check both sides, but not checking
2203 * the "old" side here is deliberate. We are checking the newly
2204 * introduced changes, and as long as the "new" side is text, we
2205 * can and should check what it introduces.
2206 */
2207 if (diff_filespec_is_binary(two))
2208 goto free_and_return;
2209 else {
2210 /* Crazy xdl interfaces.. */
2211 xpparam_t xpp;
2212 xdemitconf_t xecfg;
2213
2214 memset(&xpp, 0, sizeof(xpp));
2215 memset(&xecfg, 0, sizeof(xecfg));
2216 xecfg.ctxlen = 1; /* at least one context line */
2217 xpp.flags = 0;
2218 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2219 &xpp, &xecfg);
2220
2221 if (data.ws_rule & WS_BLANK_AT_EOF) {
2222 struct emit_callback ecbdata;
2223 int blank_at_eof;
2224
2225 ecbdata.ws_rule = data.ws_rule;
2226 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2227 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2228
2229 if (blank_at_eof) {
2230 static char *err;
2231 if (!err)
2232 err = whitespace_error_string(WS_BLANK_AT_EOF);
2233 fprintf(o->file, "%s:%d: %s.\n",
2234 data.filename, blank_at_eof, err);
2235 data.status = 1; /* report errors */
2236 }
2237 }
2238 }
2239 free_and_return:
2240 diff_free_filespec_data(one);
2241 diff_free_filespec_data(two);
2242 if (data.status)
2243 DIFF_OPT_SET(o, CHECK_FAILED);
2244 }
2245
2246 struct diff_filespec *alloc_filespec(const char *path)
2247 {
2248 int namelen = strlen(path);
2249 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2250
2251 memset(spec, 0, sizeof(*spec));
2252 spec->path = (char *)(spec + 1);
2253 memcpy(spec->path, path, namelen+1);
2254 spec->count = 1;
2255 spec->is_binary = -1;
2256 return spec;
2257 }
2258
2259 void free_filespec(struct diff_filespec *spec)
2260 {
2261 if (!--spec->count) {
2262 diff_free_filespec_data(spec);
2263 free(spec);
2264 }
2265 }
2266
2267 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2268 unsigned short mode)
2269 {
2270 if (mode) {
2271 spec->mode = canon_mode(mode);
2272 hashcpy(spec->sha1, sha1);
2273 spec->sha1_valid = !is_null_sha1(sha1);
2274 }
2275 }
2276
2277 /*
2278 * Given a name and sha1 pair, if the index tells us the file in
2279 * the work tree has that object contents, return true, so that
2280 * prepare_temp_file() does not have to inflate and extract.
2281 */
2282 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2283 {
2284 struct cache_entry *ce;
2285 struct stat st;
2286 int pos, len;
2287
2288 /*
2289 * We do not read the cache ourselves here, because the
2290 * benchmark with my previous version that always reads cache
2291 * shows that it makes things worse for diff-tree comparing
2292 * two linux-2.6 kernel trees in an already checked out work
2293 * tree. This is because most diff-tree comparisons deal with
2294 * only a small number of files, while reading the cache is
2295 * expensive for a large project, and its cost outweighs the
2296 * savings we get by not inflating the object to a temporary
2297 * file. Practically, this code only helps when we are used
2298 * by diff-cache --cached, which does read the cache before
2299 * calling us.
2300 */
2301 if (!active_cache)
2302 return 0;
2303
2304 /* We want to avoid the working directory if our caller
2305 * doesn't need the data in a normal file, this system
2306 * is rather slow with its stat/open/mmap/close syscalls,
2307 * and the object is contained in a pack file. The pack
2308 * is probably already open and will be faster to obtain
2309 * the data through than the working directory. Loose
2310 * objects however would tend to be slower as they need
2311 * to be individually opened and inflated.
2312 */
2313 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2314 return 0;
2315
2316 len = strlen(name);
2317 pos = cache_name_pos(name, len);
2318 if (pos < 0)
2319 return 0;
2320 ce = active_cache[pos];
2321
2322 /*
2323 * This is not the sha1 we are looking for, or
2324 * unreusable because it is not a regular file.
2325 */
2326 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2327 return 0;
2328
2329 /*
2330 * If ce is marked as "assume unchanged", there is no
2331 * guarantee that work tree matches what we are looking for.
2332 */
2333 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2334 return 0;
2335
2336 /*
2337 * If ce matches the file in the work tree, we can reuse it.
2338 */
2339 if (ce_uptodate(ce) ||
2340 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2341 return 1;
2342
2343 return 0;
2344 }
2345
2346 static int populate_from_stdin(struct diff_filespec *s)
2347 {
2348 struct strbuf buf = STRBUF_INIT;
2349 size_t size = 0;
2350
2351 if (strbuf_read(&buf, 0, 0) < 0)
2352 return error("error while reading from stdin %s",
2353 strerror(errno));
2354
2355 s->should_munmap = 0;
2356 s->data = strbuf_detach(&buf, &size);
2357 s->size = size;
2358 s->should_free = 1;
2359 return 0;
2360 }
2361
2362 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2363 {
2364 int len;
2365 char *data = xmalloc(100), *dirty = "";
2366
2367 /* Are we looking at the work tree? */
2368 if (s->dirty_submodule)
2369 dirty = "-dirty";
2370
2371 len = snprintf(data, 100,
2372 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2373 s->data = data;
2374 s->size = len;
2375 s->should_free = 1;
2376 if (size_only) {
2377 s->data = NULL;
2378 free(data);
2379 }
2380 return 0;
2381 }
2382
2383 /*
2384 * While doing rename detection and pickaxe operation, we may need to
2385 * grab the data for the blob (or file) for our own in-core comparison.
2386 * diff_filespec has data and size fields for this purpose.
2387 */
2388 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2389 {
2390 int err = 0;
2391 if (!DIFF_FILE_VALID(s))
2392 die("internal error: asking to populate invalid file.");
2393 if (S_ISDIR(s->mode))
2394 return -1;
2395
2396 if (s->data)
2397 return 0;
2398
2399 if (size_only && 0 < s->size)
2400 return 0;
2401
2402 if (S_ISGITLINK(s->mode))
2403 return diff_populate_gitlink(s, size_only);
2404
2405 if (!s->sha1_valid ||
2406 reuse_worktree_file(s->path, s->sha1, 0)) {
2407 struct strbuf buf = STRBUF_INIT;
2408 struct stat st;
2409 int fd;
2410
2411 if (!strcmp(s->path, "-"))
2412 return populate_from_stdin(s);
2413
2414 if (lstat(s->path, &st) < 0) {
2415 if (errno == ENOENT) {
2416 err_empty:
2417 err = -1;
2418 empty:
2419 s->data = (char *)"";
2420 s->size = 0;
2421 return err;
2422 }
2423 }
2424 s->size = xsize_t(st.st_size);
2425 if (!s->size)
2426 goto empty;
2427 if (S_ISLNK(st.st_mode)) {
2428 struct strbuf sb = STRBUF_INIT;
2429
2430 if (strbuf_readlink(&sb, s->path, s->size))
2431 goto err_empty;
2432 s->size = sb.len;
2433 s->data = strbuf_detach(&sb, NULL);
2434 s->should_free = 1;
2435 return 0;
2436 }
2437 if (size_only)
2438 return 0;
2439 fd = open(s->path, O_RDONLY);
2440 if (fd < 0)
2441 goto err_empty;
2442 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2443 close(fd);
2444 s->should_munmap = 1;
2445
2446 /*
2447 * Convert from working tree format to canonical git format
2448 */
2449 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2450 size_t size = 0;
2451 munmap(s->data, s->size);
2452 s->should_munmap = 0;
2453 s->data = strbuf_detach(&buf, &size);
2454 s->size = size;
2455 s->should_free = 1;
2456 }
2457 }
2458 else {
2459 enum object_type type;
2460 if (size_only) {
2461 type = sha1_object_info(s->sha1, &s->size);
2462 if (type < 0)
2463 die("unable to read %s", sha1_to_hex(s->sha1));
2464 } else {
2465 s->data = read_sha1_file(s->sha1, &type, &s->size);
2466 if (!s->data)
2467 die("unable to read %s", sha1_to_hex(s->sha1));
2468 s->should_free = 1;
2469 }
2470 }
2471 return 0;
2472 }
2473
2474 void diff_free_filespec_blob(struct diff_filespec *s)
2475 {
2476 if (s->should_free)
2477 free(s->data);
2478 else if (s->should_munmap)
2479 munmap(s->data, s->size);
2480
2481 if (s->should_free || s->should_munmap) {
2482 s->should_free = s->should_munmap = 0;
2483 s->data = NULL;
2484 }
2485 }
2486
2487 void diff_free_filespec_data(struct diff_filespec *s)
2488 {
2489 diff_free_filespec_blob(s);
2490 free(s->cnt_data);
2491 s->cnt_data = NULL;
2492 }
2493
2494 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2495 void *blob,
2496 unsigned long size,
2497 const unsigned char *sha1,
2498 int mode)
2499 {
2500 int fd;
2501 struct strbuf buf = STRBUF_INIT;
2502 struct strbuf template = STRBUF_INIT;
2503 char *path_dup = xstrdup(path);
2504 const char *base = basename(path_dup);
2505
2506 /* Generate "XXXXXX_basename.ext" */
2507 strbuf_addstr(&template, "XXXXXX_");
2508 strbuf_addstr(&template, base);
2509
2510 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2511 strlen(base) + 1);
2512 if (fd < 0)
2513 die_errno("unable to create temp-file");
2514 if (convert_to_working_tree(path,
2515 (const char *)blob, (size_t)size, &buf)) {
2516 blob = buf.buf;
2517 size = buf.len;
2518 }
2519 if (write_in_full(fd, blob, size) != size)
2520 die_errno("unable to write temp-file");
2521 close(fd);
2522 temp->name = temp->tmp_path;
2523 strcpy(temp->hex, sha1_to_hex(sha1));
2524 temp->hex[40] = 0;
2525 sprintf(temp->mode, "%06o", mode);
2526 strbuf_release(&buf);
2527 strbuf_release(&template);
2528 free(path_dup);
2529 }
2530
2531 static struct diff_tempfile *prepare_temp_file(const char *name,
2532 struct diff_filespec *one)
2533 {
2534 struct diff_tempfile *temp = claim_diff_tempfile();
2535
2536 if (!DIFF_FILE_VALID(one)) {
2537 not_a_valid_file:
2538 /* A '-' entry produces this for file-2, and
2539 * a '+' entry produces this for file-1.
2540 */
2541 temp->name = "/dev/null";
2542 strcpy(temp->hex, ".");
2543 strcpy(temp->mode, ".");
2544 return temp;
2545 }
2546
2547 if (!remove_tempfile_installed) {
2548 atexit(remove_tempfile);
2549 sigchain_push_common(remove_tempfile_on_signal);
2550 remove_tempfile_installed = 1;
2551 }
2552
2553 if (!one->sha1_valid ||
2554 reuse_worktree_file(name, one->sha1, 1)) {
2555 struct stat st;
2556 if (lstat(name, &st) < 0) {
2557 if (errno == ENOENT)
2558 goto not_a_valid_file;
2559 die_errno("stat(%s)", name);
2560 }
2561 if (S_ISLNK(st.st_mode)) {
2562 struct strbuf sb = STRBUF_INIT;
2563 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2564 die_errno("readlink(%s)", name);
2565 prep_temp_blob(name, temp, sb.buf, sb.len,
2566 (one->sha1_valid ?
2567 one->sha1 : null_sha1),
2568 (one->sha1_valid ?
2569 one->mode : S_IFLNK));
2570 strbuf_release(&sb);
2571 }
2572 else {
2573 /* we can borrow from the file in the work tree */
2574 temp->name = name;
2575 if (!one->sha1_valid)
2576 strcpy(temp->hex, sha1_to_hex(null_sha1));
2577 else
2578 strcpy(temp->hex, sha1_to_hex(one->sha1));
2579 /* Even though we may sometimes borrow the
2580 * contents from the work tree, we always want
2581 * one->mode. mode is trustworthy even when
2582 * !(one->sha1_valid), as long as
2583 * DIFF_FILE_VALID(one).
2584 */
2585 sprintf(temp->mode, "%06o", one->mode);
2586 }
2587 return temp;
2588 }
2589 else {
2590 if (diff_populate_filespec(one, 0))
2591 die("cannot read data blob for %s", one->path);
2592 prep_temp_blob(name, temp, one->data, one->size,
2593 one->sha1, one->mode);
2594 }
2595 return temp;
2596 }
2597
2598 /* An external diff command takes:
2599 *
2600 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2601 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2602 *
2603 */
2604 static void run_external_diff(const char *pgm,
2605 const char *name,
2606 const char *other,
2607 struct diff_filespec *one,
2608 struct diff_filespec *two,
2609 const char *xfrm_msg,
2610 int complete_rewrite)
2611 {
2612 const char *spawn_arg[10];
2613 int retval;
2614 const char **arg = &spawn_arg[0];
2615
2616 if (one && two) {
2617 struct diff_tempfile *temp_one, *temp_two;
2618 const char *othername = (other ? other : name);
2619 temp_one = prepare_temp_file(name, one);
2620 temp_two = prepare_temp_file(othername, two);
2621 *arg++ = pgm;
2622 *arg++ = name;
2623 *arg++ = temp_one->name;
2624 *arg++ = temp_one->hex;
2625 *arg++ = temp_one->mode;
2626 *arg++ = temp_two->name;
2627 *arg++ = temp_two->hex;
2628 *arg++ = temp_two->mode;
2629 if (other) {
2630 *arg++ = other;
2631 *arg++ = xfrm_msg;
2632 }
2633 } else {
2634 *arg++ = pgm;
2635 *arg++ = name;
2636 }
2637 *arg = NULL;
2638 fflush(NULL);
2639 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2640 remove_tempfile();
2641 if (retval) {
2642 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2643 exit(1);
2644 }
2645 }
2646
2647 static int similarity_index(struct diff_filepair *p)
2648 {
2649 return p->score * 100 / MAX_SCORE;
2650 }
2651
2652 static void fill_metainfo(struct strbuf *msg,
2653 const char *name,
2654 const char *other,
2655 struct diff_filespec *one,
2656 struct diff_filespec *two,
2657 struct diff_options *o,
2658 struct diff_filepair *p,
2659 int *must_show_header,
2660 int use_color)
2661 {
2662 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2663 const char *reset = diff_get_color(use_color, DIFF_RESET);
2664 struct strbuf *msgbuf;
2665 char *line_prefix = "";
2666
2667 *must_show_header = 1;
2668 if (o->output_prefix) {
2669 msgbuf = o->output_prefix(o, o->output_prefix_data);
2670 line_prefix = msgbuf->buf;
2671 }
2672 strbuf_init(msg, PATH_MAX * 2 + 300);
2673 switch (p->status) {
2674 case DIFF_STATUS_COPIED:
2675 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2676 line_prefix, set, similarity_index(p));
2677 strbuf_addf(msg, "%s\n%s%scopy from ",
2678 reset, line_prefix, set);
2679 quote_c_style(name, msg, NULL, 0);
2680 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2681 quote_c_style(other, msg, NULL, 0);
2682 strbuf_addf(msg, "%s\n", reset);
2683 break;
2684 case DIFF_STATUS_RENAMED:
2685 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2686 line_prefix, set, similarity_index(p));
2687 strbuf_addf(msg, "%s\n%s%srename from ",
2688 reset, line_prefix, set);
2689 quote_c_style(name, msg, NULL, 0);
2690 strbuf_addf(msg, "%s\n%s%srename to ",
2691 reset, line_prefix, set);
2692 quote_c_style(other, msg, NULL, 0);
2693 strbuf_addf(msg, "%s\n", reset);
2694 break;
2695 case DIFF_STATUS_MODIFIED:
2696 if (p->score) {
2697 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2698 line_prefix,
2699 set, similarity_index(p), reset);
2700 break;
2701 }
2702 /* fallthru */
2703 default:
2704 *must_show_header = 0;
2705 }
2706 if (one && two && hashcmp(one->sha1, two->sha1)) {
2707 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2708
2709 if (DIFF_OPT_TST(o, BINARY)) {
2710 mmfile_t mf;
2711 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2712 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2713 abbrev = 40;
2714 }
2715 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2716 find_unique_abbrev(one->sha1, abbrev));
2717 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2718 if (one->mode == two->mode)
2719 strbuf_addf(msg, " %06o", one->mode);
2720 strbuf_addf(msg, "%s\n", reset);
2721 }
2722 }
2723
2724 static void run_diff_cmd(const char *pgm,
2725 const char *name,
2726 const char *other,
2727 const char *attr_path,
2728 struct diff_filespec *one,
2729 struct diff_filespec *two,
2730 struct strbuf *msg,
2731 struct diff_options *o,
2732 struct diff_filepair *p)
2733 {
2734 const char *xfrm_msg = NULL;
2735 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2736 int must_show_header = 0;
2737
2738 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2739 pgm = NULL;
2740 else {
2741 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2742 if (drv && drv->external)
2743 pgm = drv->external;
2744 }
2745
2746 if (msg) {
2747 /*
2748 * don't use colors when the header is intended for an
2749 * external diff driver
2750 */
2751 fill_metainfo(msg, name, other, one, two, o, p,
2752 &must_show_header,
2753 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2754 xfrm_msg = msg->len ? msg->buf : NULL;
2755 }
2756
2757 if (pgm) {
2758 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2759 complete_rewrite);
2760 return;
2761 }
2762 if (one && two)
2763 builtin_diff(name, other ? other : name,
2764 one, two, xfrm_msg, must_show_header,
2765 o, complete_rewrite);
2766 else
2767 fprintf(o->file, "* Unmerged path %s\n", name);
2768 }
2769
2770 static void diff_fill_sha1_info(struct diff_filespec *one)
2771 {
2772 if (DIFF_FILE_VALID(one)) {
2773 if (!one->sha1_valid) {
2774 struct stat st;
2775 if (!strcmp(one->path, "-")) {
2776 hashcpy(one->sha1, null_sha1);
2777 return;
2778 }
2779 if (lstat(one->path, &st) < 0)
2780 die_errno("stat '%s'", one->path);
2781 if (index_path(one->sha1, one->path, &st, 0))
2782 die("cannot hash %s", one->path);
2783 }
2784 }
2785 else
2786 hashclr(one->sha1);
2787 }
2788
2789 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2790 {
2791 /* Strip the prefix but do not molest /dev/null and absolute paths */
2792 if (*namep && **namep != '/') {
2793 *namep += prefix_length;
2794 if (**namep == '/')
2795 ++*namep;
2796 }
2797 if (*otherp && **otherp != '/') {
2798 *otherp += prefix_length;
2799 if (**otherp == '/')
2800 ++*otherp;
2801 }
2802 }
2803
2804 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2805 {
2806 const char *pgm = external_diff();
2807 struct strbuf msg;
2808 struct diff_filespec *one = p->one;
2809 struct diff_filespec *two = p->two;
2810 const char *name;
2811 const char *other;
2812 const char *attr_path;
2813
2814 name = p->one->path;
2815 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2816 attr_path = name;
2817 if (o->prefix_length)
2818 strip_prefix(o->prefix_length, &name, &other);
2819
2820 if (DIFF_PAIR_UNMERGED(p)) {
2821 run_diff_cmd(pgm, name, NULL, attr_path,
2822 NULL, NULL, NULL, o, p);
2823 return;
2824 }
2825
2826 diff_fill_sha1_info(one);
2827 diff_fill_sha1_info(two);
2828
2829 if (!pgm &&
2830 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2831 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2832 /*
2833 * a filepair that changes between file and symlink
2834 * needs to be split into deletion and creation.
2835 */
2836 struct diff_filespec *null = alloc_filespec(two->path);
2837 run_diff_cmd(NULL, name, other, attr_path,
2838 one, null, &msg, o, p);
2839 free(null);
2840 strbuf_release(&msg);
2841
2842 null = alloc_filespec(one->path);
2843 run_diff_cmd(NULL, name, other, attr_path,
2844 null, two, &msg, o, p);
2845 free(null);
2846 }
2847 else
2848 run_diff_cmd(pgm, name, other, attr_path,
2849 one, two, &msg, o, p);
2850
2851 strbuf_release(&msg);
2852 }
2853
2854 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2855 struct diffstat_t *diffstat)
2856 {
2857 const char *name;
2858 const char *other;
2859 int complete_rewrite = 0;
2860
2861 if (DIFF_PAIR_UNMERGED(p)) {
2862 /* unmerged */
2863 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2864 return;
2865 }
2866
2867 name = p->one->path;
2868 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2869
2870 if (o->prefix_length)
2871 strip_prefix(o->prefix_length, &name, &other);
2872
2873 diff_fill_sha1_info(p->one);
2874 diff_fill_sha1_info(p->two);
2875
2876 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2877 complete_rewrite = 1;
2878 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2879 }
2880
2881 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2882 {
2883 const char *name;
2884 const char *other;
2885 const char *attr_path;
2886
2887 if (DIFF_PAIR_UNMERGED(p)) {
2888 /* unmerged */
2889 return;
2890 }
2891
2892 name = p->one->path;
2893 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2894 attr_path = other ? other : name;
2895
2896 if (o->prefix_length)
2897 strip_prefix(o->prefix_length, &name, &other);
2898
2899 diff_fill_sha1_info(p->one);
2900 diff_fill_sha1_info(p->two);
2901
2902 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2903 }
2904
2905 void diff_setup(struct diff_options *options)
2906 {
2907 memcpy(options, &default_diff_options, sizeof(*options));
2908
2909 options->file = stdout;
2910
2911 options->line_termination = '\n';
2912 options->break_opt = -1;
2913 options->rename_limit = -1;
2914 options->dirstat_percent = 3;
2915 options->context = 3;
2916
2917 options->change = diff_change;
2918 options->add_remove = diff_addremove;
2919 if (diff_use_color_default > 0)
2920 DIFF_OPT_SET(options, COLOR_DIFF);
2921 options->detect_rename = diff_detect_rename_default;
2922
2923 if (diff_no_prefix) {
2924 options->a_prefix = options->b_prefix = "";
2925 } else if (!diff_mnemonic_prefix) {
2926 options->a_prefix = "a/";
2927 options->b_prefix = "b/";
2928 }
2929 }
2930
2931 int diff_setup_done(struct diff_options *options)
2932 {
2933 int count = 0;
2934
2935 if (options->output_format & DIFF_FORMAT_NAME)
2936 count++;
2937 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2938 count++;
2939 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2940 count++;
2941 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2942 count++;
2943 if (count > 1)
2944 die("--name-only, --name-status, --check and -s are mutually exclusive");
2945
2946 /*
2947 * Most of the time we can say "there are changes"
2948 * only by checking if there are changed paths, but
2949 * --ignore-whitespace* options force us to look
2950 * inside contents.
2951 */
2952
2953 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2954 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2955 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2956 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2957 else
2958 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2959
2960 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2961 options->detect_rename = DIFF_DETECT_COPY;
2962
2963 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2964 options->prefix = NULL;
2965 if (options->prefix)
2966 options->prefix_length = strlen(options->prefix);
2967 else
2968 options->prefix_length = 0;
2969
2970 if (options->output_format & (DIFF_FORMAT_NAME |
2971 DIFF_FORMAT_NAME_STATUS |
2972 DIFF_FORMAT_CHECKDIFF |
2973 DIFF_FORMAT_NO_OUTPUT))
2974 options->output_format &= ~(DIFF_FORMAT_RAW |
2975 DIFF_FORMAT_NUMSTAT |
2976 DIFF_FORMAT_DIFFSTAT |
2977 DIFF_FORMAT_SHORTSTAT |
2978 DIFF_FORMAT_DIRSTAT |
2979 DIFF_FORMAT_SUMMARY |
2980 DIFF_FORMAT_PATCH);
2981
2982 /*
2983 * These cases always need recursive; we do not drop caller-supplied
2984 * recursive bits for other formats here.
2985 */
2986 if (options->output_format & (DIFF_FORMAT_PATCH |
2987 DIFF_FORMAT_NUMSTAT |
2988 DIFF_FORMAT_DIFFSTAT |
2989 DIFF_FORMAT_SHORTSTAT |
2990 DIFF_FORMAT_DIRSTAT |
2991 DIFF_FORMAT_SUMMARY |
2992 DIFF_FORMAT_CHECKDIFF))
2993 DIFF_OPT_SET(options, RECURSIVE);
2994 /*
2995 * Also pickaxe would not work very well if you do not say recursive
2996 */
2997 if (options->pickaxe)
2998 DIFF_OPT_SET(options, RECURSIVE);
2999 /*
3000 * When patches are generated, submodules diffed against the work tree
3001 * must be checked for dirtiness too so it can be shown in the output
3002 */
3003 if (options->output_format & DIFF_FORMAT_PATCH)
3004 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3005
3006 if (options->detect_rename && options->rename_limit < 0)
3007 options->rename_limit = diff_rename_limit_default;
3008 if (options->setup & DIFF_SETUP_USE_CACHE) {
3009 if (!active_cache)
3010 /* read-cache does not die even when it fails
3011 * so it is safe for us to do this here. Also
3012 * it does not smudge active_cache or active_nr
3013 * when it fails, so we do not have to worry about
3014 * cleaning it up ourselves either.
3015 */
3016 read_cache();
3017 }
3018 if (options->abbrev <= 0 || 40 < options->abbrev)
3019 options->abbrev = 40; /* full */
3020
3021 /*
3022 * It does not make sense to show the first hit we happened
3023 * to have found. It does not make sense not to return with
3024 * exit code in such a case either.
3025 */
3026 if (DIFF_OPT_TST(options, QUICK)) {
3027 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3028 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3029 }
3030
3031 return 0;
3032 }
3033
3034 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3035 {
3036 char c, *eq;
3037 int len;
3038
3039 if (*arg != '-')
3040 return 0;
3041 c = *++arg;
3042 if (!c)
3043 return 0;
3044 if (c == arg_short) {
3045 c = *++arg;
3046 if (!c)
3047 return 1;
3048 if (val && isdigit(c)) {
3049 char *end;
3050 int n = strtoul(arg, &end, 10);
3051 if (*end)
3052 return 0;
3053 *val = n;
3054 return 1;
3055 }
3056 return 0;
3057 }
3058 if (c != '-')
3059 return 0;
3060 arg++;
3061 eq = strchr(arg, '=');
3062 if (eq)
3063 len = eq - arg;
3064 else
3065 len = strlen(arg);
3066 if (!len || strncmp(arg, arg_long, len))
3067 return 0;
3068 if (eq) {
3069 int n;
3070 char *end;
3071 if (!isdigit(*++eq))
3072 return 0;
3073 n = strtoul(eq, &end, 10);
3074 if (*end)
3075 return 0;
3076 *val = n;
3077 }
3078 return 1;
3079 }
3080
3081 static int diff_scoreopt_parse(const char *opt);
3082
3083 static inline int short_opt(char opt, const char **argv,
3084 const char **optarg)
3085 {
3086 const char *arg = argv[0];
3087 if (arg[0] != '-' || arg[1] != opt)
3088 return 0;
3089 if (arg[2] != '\0') {
3090 *optarg = arg + 2;
3091 return 1;
3092 }
3093 if (!argv[1])
3094 die("Option '%c' requires a value", opt);
3095 *optarg = argv[1];
3096 return 2;
3097 }
3098
3099 int parse_long_opt(const char *opt, const char **argv,
3100 const char **optarg)
3101 {
3102 const char *arg = argv[0];
3103 if (arg[0] != '-' || arg[1] != '-')
3104 return 0;
3105 arg += strlen("--");
3106 if (prefixcmp(arg, opt))
3107 return 0;
3108 arg += strlen(opt);
3109 if (*arg == '=') { /* sticked form: --option=value */
3110 *optarg = arg + 1;
3111 return 1;
3112 }
3113 if (*arg != '\0')
3114 return 0;
3115 /* separate form: --option value */
3116 if (!argv[1])
3117 die("Option '--%s' requires a value", opt);
3118 *optarg = argv[1];
3119 return 2;
3120 }
3121
3122 static int stat_opt(struct diff_options *options, const char **av)
3123 {
3124 const char *arg = av[0];
3125 char *end;
3126 int width = options->stat_width;
3127 int name_width = options->stat_name_width;
3128 int count = options->stat_count;
3129 int argcount = 1;
3130
3131 arg += strlen("--stat");
3132 end = (char *)arg;
3133
3134 switch (*arg) {
3135 case '-':
3136 if (!prefixcmp(arg, "-width")) {
3137 arg += strlen("-width");
3138 if (*arg == '=')
3139 width = strtoul(arg + 1, &end, 10);
3140 else if (!*arg && !av[1])
3141 die("Option '--stat-width' requires a value");
3142 else if (!*arg) {
3143 width = strtoul(av[1], &end, 10);
3144 argcount = 2;
3145 }
3146 } else if (!prefixcmp(arg, "-name-width")) {
3147 arg += strlen("-name-width");
3148 if (*arg == '=')
3149 name_width = strtoul(arg + 1, &end, 10);
3150 else if (!*arg && !av[1])
3151 die("Option '--stat-name-width' requires a value");
3152 else if (!*arg) {
3153 name_width = strtoul(av[1], &end, 10);
3154 argcount = 2;
3155 }
3156 } else if (!prefixcmp(arg, "-count")) {
3157 arg += strlen("-count");
3158 if (*arg == '=')
3159 count = strtoul(arg + 1, &end, 10);
3160 else if (!*arg && !av[1])
3161 die("Option '--stat-count' requires a value");
3162 else if (!*arg) {
3163 count = strtoul(av[1], &end, 10);
3164 argcount = 2;
3165 }
3166 }
3167 break;
3168 case '=':
3169 width = strtoul(arg+1, &end, 10);
3170 if (*end == ',')
3171 name_width = strtoul(end+1, &end, 10);
3172 if (*end == ',')
3173 count = strtoul(end+1, &end, 10);
3174 }
3175
3176 /* Important! This checks all the error cases! */
3177 if (*end)
3178 return 0;
3179 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3180 options->stat_name_width = name_width;
3181 options->stat_width = width;
3182 options->stat_count = count;
3183 return argcount;
3184 }
3185
3186 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3187 {
3188 const char *arg = av[0];
3189 const char *optarg;
3190 int argcount;
3191
3192 /* Output format options */
3193 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3194 options->output_format |= DIFF_FORMAT_PATCH;
3195 else if (opt_arg(arg, 'U', "unified", &options->context))
3196 options->output_format |= DIFF_FORMAT_PATCH;
3197 else if (!strcmp(arg, "--raw"))
3198 options->output_format |= DIFF_FORMAT_RAW;
3199 else if (!strcmp(arg, "--patch-with-raw"))
3200 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3201 else if (!strcmp(arg, "--numstat"))
3202 options->output_format |= DIFF_FORMAT_NUMSTAT;
3203 else if (!strcmp(arg, "--shortstat"))
3204 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3205 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3206 options->output_format |= DIFF_FORMAT_DIRSTAT;
3207 else if (!strcmp(arg, "--cumulative")) {
3208 options->output_format |= DIFF_FORMAT_DIRSTAT;
3209 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3210 } else if (opt_arg(arg, 0, "dirstat-by-file",
3211 &options->dirstat_percent)) {
3212 options->output_format |= DIFF_FORMAT_DIRSTAT;
3213 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3214 }
3215 else if (!strcmp(arg, "--check"))
3216 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3217 else if (!strcmp(arg, "--summary"))
3218 options->output_format |= DIFF_FORMAT_SUMMARY;
3219 else if (!strcmp(arg, "--patch-with-stat"))
3220 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3221 else if (!strcmp(arg, "--name-only"))
3222 options->output_format |= DIFF_FORMAT_NAME;
3223 else if (!strcmp(arg, "--name-status"))
3224 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3225 else if (!strcmp(arg, "-s"))
3226 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3227 else if (!prefixcmp(arg, "--stat"))
3228 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3229 return stat_opt(options, av);
3230
3231 /* renames options */
3232 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3233 !strcmp(arg, "--break-rewrites")) {
3234 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3235 return error("invalid argument to -B: %s", arg+2);
3236 }
3237 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3238 !strcmp(arg, "--find-renames")) {
3239 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3240 return error("invalid argument to -M: %s", arg+2);
3241 options->detect_rename = DIFF_DETECT_RENAME;
3242 }
3243 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3244 options->irreversible_delete = 1;
3245 }
3246 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3247 !strcmp(arg, "--find-copies")) {
3248 if (options->detect_rename == DIFF_DETECT_COPY)
3249 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3250 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3251 return error("invalid argument to -C: %s", arg+2);
3252 options->detect_rename = DIFF_DETECT_COPY;
3253 }
3254 else if (!strcmp(arg, "--no-renames"))
3255 options->detect_rename = 0;
3256 else if (!strcmp(arg, "--relative"))
3257 DIFF_OPT_SET(options, RELATIVE_NAME);
3258 else if (!prefixcmp(arg, "--relative=")) {
3259 DIFF_OPT_SET(options, RELATIVE_NAME);
3260 options->prefix = arg + 11;
3261 }
3262
3263 /* xdiff options */
3264 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3265 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3266 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3267 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3268 else if (!strcmp(arg, "--ignore-space-at-eol"))
3269 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3270 else if (!strcmp(arg, "--patience"))
3271 DIFF_XDL_SET(options, PATIENCE_DIFF);
3272
3273 /* flags options */
3274 else if (!strcmp(arg, "--binary")) {
3275 options->output_format |= DIFF_FORMAT_PATCH;
3276 DIFF_OPT_SET(options, BINARY);
3277 }
3278 else if (!strcmp(arg, "--full-index"))
3279 DIFF_OPT_SET(options, FULL_INDEX);
3280 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3281 DIFF_OPT_SET(options, TEXT);
3282 else if (!strcmp(arg, "-R"))
3283 DIFF_OPT_SET(options, REVERSE_DIFF);
3284 else if (!strcmp(arg, "--find-copies-harder"))
3285 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3286 else if (!strcmp(arg, "--follow"))
3287 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3288 else if (!strcmp(arg, "--color"))
3289 DIFF_OPT_SET(options, COLOR_DIFF);
3290 else if (!prefixcmp(arg, "--color=")) {
3291 int value = git_config_colorbool(NULL, arg+8, -1);
3292 if (value == 0)
3293 DIFF_OPT_CLR(options, COLOR_DIFF);
3294 else if (value > 0)
3295 DIFF_OPT_SET(options, COLOR_DIFF);
3296 else
3297 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3298 }
3299 else if (!strcmp(arg, "--no-color"))
3300 DIFF_OPT_CLR(options, COLOR_DIFF);
3301 else if (!strcmp(arg, "--color-words")) {
3302 DIFF_OPT_SET(options, COLOR_DIFF);
3303 options->word_diff = DIFF_WORDS_COLOR;
3304 }
3305 else if (!prefixcmp(arg, "--color-words=")) {
3306 DIFF_OPT_SET(options, COLOR_DIFF);
3307 options->word_diff = DIFF_WORDS_COLOR;
3308 options->word_regex = arg + 14;
3309 }
3310 else if (!strcmp(arg, "--word-diff")) {
3311 if (options->word_diff == DIFF_WORDS_NONE)
3312 options->word_diff = DIFF_WORDS_PLAIN;
3313 }
3314 else if (!prefixcmp(arg, "--word-diff=")) {
3315 const char *type = arg + 12;
3316 if (!strcmp(type, "plain"))
3317 options->word_diff = DIFF_WORDS_PLAIN;
3318 else if (!strcmp(type, "color")) {
3319 DIFF_OPT_SET(options, COLOR_DIFF);
3320 options->word_diff = DIFF_WORDS_COLOR;
3321 }
3322 else if (!strcmp(type, "porcelain"))
3323 options->word_diff = DIFF_WORDS_PORCELAIN;
3324 else if (!strcmp(type, "none"))
3325 options->word_diff = DIFF_WORDS_NONE;
3326 else
3327 die("bad --word-diff argument: %s", type);
3328 }
3329 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3330 if (options->word_diff == DIFF_WORDS_NONE)
3331 options->word_diff = DIFF_WORDS_PLAIN;
3332 options->word_regex = optarg;
3333 return argcount;
3334 }
3335 else if (!strcmp(arg, "--exit-code"))
3336 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3337 else if (!strcmp(arg, "--quiet"))
3338 DIFF_OPT_SET(options, QUICK);
3339 else if (!strcmp(arg, "--ext-diff"))
3340 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3341 else if (!strcmp(arg, "--no-ext-diff"))
3342 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3343 else if (!strcmp(arg, "--textconv"))
3344 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3345 else if (!strcmp(arg, "--no-textconv"))
3346 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3347 else if (!strcmp(arg, "--ignore-submodules")) {
3348 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3349 handle_ignore_submodules_arg(options, "all");
3350 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3351 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3352 handle_ignore_submodules_arg(options, arg + 20);
3353 } else if (!strcmp(arg, "--submodule"))
3354 DIFF_OPT_SET(options, SUBMODULE_LOG);
3355 else if (!prefixcmp(arg, "--submodule=")) {
3356 if (!strcmp(arg + 12, "log"))
3357 DIFF_OPT_SET(options, SUBMODULE_LOG);
3358 }
3359
3360 /* misc options */
3361 else if (!strcmp(arg, "-z"))
3362 options->line_termination = 0;
3363 else if ((argcount = short_opt('l', av, &optarg))) {
3364 options->rename_limit = strtoul(optarg, NULL, 10);
3365 return argcount;
3366 }
3367 else if ((argcount = short_opt('S', av, &optarg))) {
3368 options->pickaxe = optarg;
3369 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3370 return argcount;
3371 } else if ((argcount = short_opt('G', av, &optarg))) {
3372 options->pickaxe = optarg;
3373 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3374 return argcount;
3375 }
3376 else if (!strcmp(arg, "--pickaxe-all"))
3377 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3378 else if (!strcmp(arg, "--pickaxe-regex"))
3379 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3380 else if ((argcount = short_opt('O', av, &optarg))) {
3381 options->orderfile = optarg;
3382 return argcount;
3383 }
3384 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3385 options->filter = optarg;
3386 return argcount;
3387 }
3388 else if (!strcmp(arg, "--abbrev"))
3389 options->abbrev = DEFAULT_ABBREV;
3390 else if (!prefixcmp(arg, "--abbrev=")) {
3391 options->abbrev = strtoul(arg + 9, NULL, 10);
3392 if (options->abbrev < MINIMUM_ABBREV)
3393 options->abbrev = MINIMUM_ABBREV;
3394 else if (40 < options->abbrev)
3395 options->abbrev = 40;
3396 }
3397 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3398 options->a_prefix = optarg;
3399 return argcount;
3400 }
3401 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3402 options->b_prefix = optarg;
3403 return argcount;
3404 }
3405 else if (!strcmp(arg, "--no-prefix"))
3406 options->a_prefix = options->b_prefix = "";
3407 else if (opt_arg(arg, '\0', "inter-hunk-context",
3408 &options->interhunkcontext))
3409 ;
3410 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3411 options->file = fopen(optarg, "w");
3412 if (!options->file)
3413 die_errno("Could not open '%s'", optarg);
3414 options->close_file = 1;
3415 return argcount;
3416 } else
3417 return 0;
3418 return 1;
3419 }
3420
3421 int parse_rename_score(const char **cp_p)
3422 {
3423 unsigned long num, scale;
3424 int ch, dot;
3425 const char *cp = *cp_p;
3426
3427 num = 0;
3428 scale = 1;
3429 dot = 0;
3430 for (;;) {
3431 ch = *cp;
3432 if ( !dot && ch == '.' ) {
3433 scale = 1;
3434 dot = 1;
3435 } else if ( ch == '%' ) {
3436 scale = dot ? scale*100 : 100;
3437 cp++; /* % is always at the end */
3438 break;
3439 } else if ( ch >= '0' && ch <= '9' ) {
3440 if ( scale < 100000 ) {
3441 scale *= 10;
3442 num = (num*10) + (ch-'0');
3443 }
3444 } else {
3445 break;
3446 }
3447 cp++;
3448 }
3449 *cp_p = cp;
3450
3451 /* user says num divided by scale and we say internally that
3452 * is MAX_SCORE * num / scale.
3453 */
3454 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3455 }
3456
3457 static int diff_scoreopt_parse(const char *opt)
3458 {
3459 int opt1, opt2, cmd;
3460
3461 if (*opt++ != '-')
3462 return -1;
3463 cmd = *opt++;
3464 if (cmd == '-') {
3465 /* convert the long-form arguments into short-form versions */
3466 if (!prefixcmp(opt, "break-rewrites")) {
3467 opt += strlen("break-rewrites");
3468 if (*opt == 0 || *opt++ == '=')
3469 cmd = 'B';
3470 } else if (!prefixcmp(opt, "find-copies")) {
3471 opt += strlen("find-copies");
3472 if (*opt == 0 || *opt++ == '=')
3473 cmd = 'C';
3474 } else if (!prefixcmp(opt, "find-renames")) {
3475 opt += strlen("find-renames");
3476 if (*opt == 0 || *opt++ == '=')
3477 cmd = 'M';
3478 }
3479 }
3480 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3481 return -1; /* that is not a -M, -C nor -B option */
3482
3483 opt1 = parse_rename_score(&opt);
3484 if (cmd != 'B')
3485 opt2 = 0;
3486 else {
3487 if (*opt == 0)
3488 opt2 = 0;
3489 else if (*opt != '/')
3490 return -1; /* we expect -B80/99 or -B80 */
3491 else {
3492 opt++;
3493 opt2 = parse_rename_score(&opt);
3494 }
3495 }
3496 if (*opt != 0)
3497 return -1;
3498 return opt1 | (opt2 << 16);
3499 }
3500
3501 struct diff_queue_struct diff_queued_diff;
3502
3503 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3504 {
3505 if (queue->alloc <= queue->nr) {
3506 queue->alloc = alloc_nr(queue->alloc);
3507 queue->queue = xrealloc(queue->queue,
3508 sizeof(dp) * queue->alloc);
3509 }
3510 queue->queue[queue->nr++] = dp;
3511 }
3512
3513 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3514 struct diff_filespec *one,
3515 struct diff_filespec *two)
3516 {
3517 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3518 dp->one = one;
3519 dp->two = two;
3520 if (queue)
3521 diff_q(queue, dp);
3522 return dp;
3523 }
3524
3525 void diff_free_filepair(struct diff_filepair *p)
3526 {
3527 free_filespec(p->one);
3528 free_filespec(p->two);
3529 free(p);
3530 }
3531
3532 /* This is different from find_unique_abbrev() in that
3533 * it stuffs the result with dots for alignment.
3534 */
3535 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3536 {
3537 int abblen;
3538 const char *abbrev;
3539 if (len == 40)
3540 return sha1_to_hex(sha1);
3541
3542 abbrev = find_unique_abbrev(sha1, len);
3543 abblen = strlen(abbrev);
3544 if (abblen < 37) {
3545 static char hex[41];
3546 if (len < abblen && abblen <= len + 2)
3547 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3548 else
3549 sprintf(hex, "%s...", abbrev);
3550 return hex;
3551 }
3552 return sha1_to_hex(sha1);
3553 }
3554
3555 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3556 {
3557 int line_termination = opt->line_termination;
3558 int inter_name_termination = line_termination ? '\t' : '\0';
3559 if (opt->output_prefix) {
3560 struct strbuf *msg = NULL;
3561 msg = opt->output_prefix(opt, opt->output_prefix_data);
3562 fprintf(opt->file, "%s", msg->buf);
3563 }
3564
3565 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3566 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3567 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3568 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3569 }
3570 if (p->score) {
3571 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3572 inter_name_termination);
3573 } else {
3574 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3575 }
3576
3577 if (p->status == DIFF_STATUS_COPIED ||
3578 p->status == DIFF_STATUS_RENAMED) {
3579 const char *name_a, *name_b;
3580 name_a = p->one->path;
3581 name_b = p->two->path;
3582 strip_prefix(opt->prefix_length, &name_a, &name_b);
3583 write_name_quoted(name_a, opt->file, inter_name_termination);
3584 write_name_quoted(name_b, opt->file, line_termination);
3585 } else {
3586 const char *name_a, *name_b;
3587 name_a = p->one->mode ? p->one->path : p->two->path;
3588 name_b = NULL;
3589 strip_prefix(opt->prefix_length, &name_a, &name_b);
3590 write_name_quoted(name_a, opt->file, line_termination);
3591 }
3592 }
3593
3594 int diff_unmodified_pair(struct diff_filepair *p)
3595 {
3596 /* This function is written stricter than necessary to support
3597 * the currently implemented transformers, but the idea is to
3598 * let transformers to produce diff_filepairs any way they want,
3599 * and filter and clean them up here before producing the output.
3600 */
3601 struct diff_filespec *one = p->one, *two = p->two;
3602
3603 if (DIFF_PAIR_UNMERGED(p))
3604 return 0; /* unmerged is interesting */
3605
3606 /* deletion, addition, mode or type change
3607 * and rename are all interesting.
3608 */
3609 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3610 DIFF_PAIR_MODE_CHANGED(p) ||
3611 strcmp(one->path, two->path))
3612 return 0;
3613
3614 /* both are valid and point at the same path. that is, we are
3615 * dealing with a change.
3616 */
3617 if (one->sha1_valid && two->sha1_valid &&
3618 !hashcmp(one->sha1, two->sha1) &&
3619 !one->dirty_submodule && !two->dirty_submodule)
3620 return 1; /* no change */
3621 if (!one->sha1_valid && !two->sha1_valid)
3622 return 1; /* both look at the same file on the filesystem. */
3623 return 0;
3624 }
3625
3626 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3627 {
3628 if (diff_unmodified_pair(p))
3629 return;
3630
3631 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3632 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3633 return; /* no tree diffs in patch format */
3634
3635 run_diff(p, o);
3636 }
3637
3638 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3639 struct diffstat_t *diffstat)
3640 {
3641 if (diff_unmodified_pair(p))
3642 return;
3643
3644 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3645 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3646 return; /* no useful stat for tree diffs */
3647
3648 run_diffstat(p, o, diffstat);
3649 }
3650
3651 static void diff_flush_checkdiff(struct diff_filepair *p,
3652 struct diff_options *o)
3653 {
3654 if (diff_unmodified_pair(p))
3655 return;
3656
3657 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3658 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3659 return; /* nothing to check in tree diffs */
3660
3661 run_checkdiff(p, o);
3662 }
3663
3664 int diff_queue_is_empty(void)
3665 {
3666 struct diff_queue_struct *q = &diff_queued_diff;
3667 int i;
3668 for (i = 0; i < q->nr; i++)
3669 if (!diff_unmodified_pair(q->queue[i]))
3670 return 0;
3671 return 1;
3672 }
3673
3674 #if DIFF_DEBUG
3675 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3676 {
3677 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3678 x, one ? one : "",
3679 s->path,
3680 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3681 s->mode,
3682 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3683 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3684 x, one ? one : "",
3685 s->size, s->xfrm_flags);
3686 }
3687
3688 void diff_debug_filepair(const struct diff_filepair *p, int i)
3689 {
3690 diff_debug_filespec(p->one, i, "one");
3691 diff_debug_filespec(p->two, i, "two");
3692 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3693 p->score, p->status ? p->status : '?',
3694 p->one->rename_used, p->broken_pair);
3695 }
3696
3697 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3698 {
3699 int i;
3700 if (msg)
3701 fprintf(stderr, "%s\n", msg);
3702 fprintf(stderr, "q->nr = %d\n", q->nr);
3703 for (i = 0; i < q->nr; i++) {
3704 struct diff_filepair *p = q->queue[i];
3705 diff_debug_filepair(p, i);
3706 }
3707 }
3708 #endif
3709
3710 static void diff_resolve_rename_copy(void)
3711 {
3712 int i;
3713 struct diff_filepair *p;
3714 struct diff_queue_struct *q = &diff_queued_diff;
3715
3716 diff_debug_queue("resolve-rename-copy", q);
3717
3718 for (i = 0; i < q->nr; i++) {
3719 p = q->queue[i];
3720 p->status = 0; /* undecided */
3721 if (DIFF_PAIR_UNMERGED(p))
3722 p->status = DIFF_STATUS_UNMERGED;
3723 else if (!DIFF_FILE_VALID(p->one))
3724 p->status = DIFF_STATUS_ADDED;
3725 else if (!DIFF_FILE_VALID(p->two))
3726 p->status = DIFF_STATUS_DELETED;
3727 else if (DIFF_PAIR_TYPE_CHANGED(p))
3728 p->status = DIFF_STATUS_TYPE_CHANGED;
3729
3730 /* from this point on, we are dealing with a pair
3731 * whose both sides are valid and of the same type, i.e.
3732 * either in-place edit or rename/copy edit.
3733 */
3734 else if (DIFF_PAIR_RENAME(p)) {
3735 /*
3736 * A rename might have re-connected a broken
3737 * pair up, causing the pathnames to be the
3738 * same again. If so, that's not a rename at
3739 * all, just a modification..
3740 *
3741 * Otherwise, see if this source was used for
3742 * multiple renames, in which case we decrement
3743 * the count, and call it a copy.
3744 */
3745 if (!strcmp(p->one->path, p->two->path))
3746 p->status = DIFF_STATUS_MODIFIED;
3747 else if (--p->one->rename_used > 0)
3748 p->status = DIFF_STATUS_COPIED;
3749 else
3750 p->status = DIFF_STATUS_RENAMED;
3751 }
3752 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3753 p->one->mode != p->two->mode ||
3754 p->one->dirty_submodule ||
3755 p->two->dirty_submodule ||
3756 is_null_sha1(p->one->sha1))
3757 p->status = DIFF_STATUS_MODIFIED;
3758 else {
3759 /* This is a "no-change" entry and should not
3760 * happen anymore, but prepare for broken callers.
3761 */
3762 error("feeding unmodified %s to diffcore",
3763 p->one->path);
3764 p->status = DIFF_STATUS_UNKNOWN;
3765 }
3766 }
3767 diff_debug_queue("resolve-rename-copy done", q);
3768 }
3769
3770 static int check_pair_status(struct diff_filepair *p)
3771 {
3772 switch (p->status) {
3773 case DIFF_STATUS_UNKNOWN:
3774 return 0;
3775 case 0:
3776 die("internal error in diff-resolve-rename-copy");
3777 default:
3778 return 1;
3779 }
3780 }
3781
3782 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3783 {
3784 int fmt = opt->output_format;
3785
3786 if (fmt & DIFF_FORMAT_CHECKDIFF)
3787 diff_flush_checkdiff(p, opt);
3788 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3789 diff_flush_raw(p, opt);
3790 else if (fmt & DIFF_FORMAT_NAME) {
3791 const char *name_a, *name_b;
3792 name_a = p->two->path;
3793 name_b = NULL;
3794 strip_prefix(opt->prefix_length, &name_a, &name_b);
3795 write_name_quoted(name_a, opt->file, opt->line_termination);
3796 }
3797 }
3798
3799 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3800 {
3801 if (fs->mode)
3802 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3803 else
3804 fprintf(file, " %s ", newdelete);
3805 write_name_quoted(fs->path, file, '\n');
3806 }
3807
3808
3809 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3810 const char *line_prefix)
3811 {
3812 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3813 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3814 p->two->mode, show_name ? ' ' : '\n');
3815 if (show_name) {
3816 write_name_quoted(p->two->path, file, '\n');
3817 }
3818 }
3819 }
3820
3821 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3822 const char *line_prefix)
3823 {
3824 char *names = pprint_rename(p->one->path, p->two->path);
3825
3826 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3827 free(names);
3828 show_mode_change(file, p, 0, line_prefix);
3829 }
3830
3831 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3832 {
3833 FILE *file = opt->file;
3834 char *line_prefix = "";
3835
3836 if (opt->output_prefix) {
3837 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3838 line_prefix = buf->buf;
3839 }
3840
3841 switch(p->status) {
3842 case DIFF_STATUS_DELETED:
3843 fputs(line_prefix, file);
3844 show_file_mode_name(file, "delete", p->one);
3845 break;
3846 case DIFF_STATUS_ADDED:
3847 fputs(line_prefix, file);
3848 show_file_mode_name(file, "create", p->two);
3849 break;
3850 case DIFF_STATUS_COPIED:
3851 fputs(line_prefix, file);
3852 show_rename_copy(file, "copy", p, line_prefix);
3853 break;
3854 case DIFF_STATUS_RENAMED:
3855 fputs(line_prefix, file);
3856 show_rename_copy(file, "rename", p, line_prefix);
3857 break;
3858 default:
3859 if (p->score) {
3860 fprintf(file, "%s rewrite ", line_prefix);
3861 write_name_quoted(p->two->path, file, ' ');
3862 fprintf(file, "(%d%%)\n", similarity_index(p));
3863 }
3864 show_mode_change(file, p, !p->score, line_prefix);
3865 break;
3866 }
3867 }
3868
3869 struct patch_id_t {
3870 git_SHA_CTX *ctx;
3871 int patchlen;
3872 };
3873
3874 static int remove_space(char *line, int len)
3875 {
3876 int i;
3877 char *dst = line;
3878 unsigned char c;
3879
3880 for (i = 0; i < len; i++)
3881 if (!isspace((c = line[i])))
3882 *dst++ = c;
3883
3884 return dst - line;
3885 }
3886
3887 static void patch_id_consume(void *priv, char *line, unsigned long len)
3888 {
3889 struct patch_id_t *data = priv;
3890 int new_len;
3891
3892 /* Ignore line numbers when computing the SHA1 of the patch */
3893 if (!prefixcmp(line, "@@ -"))
3894 return;
3895
3896 new_len = remove_space(line, len);
3897
3898 git_SHA1_Update(data->ctx, line, new_len);
3899 data->patchlen += new_len;
3900 }
3901
3902 /* returns 0 upon success, and writes result into sha1 */
3903 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3904 {
3905 struct diff_queue_struct *q = &diff_queued_diff;
3906 int i;
3907 git_SHA_CTX ctx;
3908 struct patch_id_t data;
3909 char buffer[PATH_MAX * 4 + 20];
3910
3911 git_SHA1_Init(&ctx);
3912 memset(&data, 0, sizeof(struct patch_id_t));
3913 data.ctx = &ctx;
3914
3915 for (i = 0; i < q->nr; i++) {
3916 xpparam_t xpp;
3917 xdemitconf_t xecfg;
3918 mmfile_t mf1, mf2;
3919 struct diff_filepair *p = q->queue[i];
3920 int len1, len2;
3921
3922 memset(&xpp, 0, sizeof(xpp));
3923 memset(&xecfg, 0, sizeof(xecfg));
3924 if (p->status == 0)
3925 return error("internal diff status error");
3926 if (p->status == DIFF_STATUS_UNKNOWN)
3927 continue;
3928 if (diff_unmodified_pair(p))
3929 continue;
3930 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3931 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3932 continue;
3933 if (DIFF_PAIR_UNMERGED(p))
3934 continue;
3935
3936 diff_fill_sha1_info(p->one);
3937 diff_fill_sha1_info(p->two);
3938 if (fill_mmfile(&mf1, p->one) < 0 ||
3939 fill_mmfile(&mf2, p->two) < 0)
3940 return error("unable to read files to diff");
3941
3942 len1 = remove_space(p->one->path, strlen(p->one->path));
3943 len2 = remove_space(p->two->path, strlen(p->two->path));
3944 if (p->one->mode == 0)
3945 len1 = snprintf(buffer, sizeof(buffer),
3946 "diff--gita/%.*sb/%.*s"
3947 "newfilemode%06o"
3948 "---/dev/null"
3949 "+++b/%.*s",
3950 len1, p->one->path,
3951 len2, p->two->path,
3952 p->two->mode,
3953 len2, p->two->path);
3954 else if (p->two->mode == 0)
3955 len1 = snprintf(buffer, sizeof(buffer),
3956 "diff--gita/%.*sb/%.*s"
3957 "deletedfilemode%06o"
3958 "---a/%.*s"
3959 "+++/dev/null",
3960 len1, p->one->path,
3961 len2, p->two->path,
3962 p->one->mode,
3963 len1, p->one->path);
3964 else
3965 len1 = snprintf(buffer, sizeof(buffer),
3966 "diff--gita/%.*sb/%.*s"
3967 "---a/%.*s"
3968 "+++b/%.*s",
3969 len1, p->one->path,
3970 len2, p->two->path,
3971 len1, p->one->path,
3972 len2, p->two->path);
3973 git_SHA1_Update(&ctx, buffer, len1);
3974
3975 if (diff_filespec_is_binary(p->one) ||
3976 diff_filespec_is_binary(p->two)) {
3977 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3978 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3979 continue;
3980 }
3981
3982 xpp.flags = 0;
3983 xecfg.ctxlen = 3;
3984 xecfg.flags = 0;
3985 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3986 &xpp, &xecfg);
3987 }
3988
3989 git_SHA1_Final(sha1, &ctx);
3990 return 0;
3991 }
3992
3993 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3994 {
3995 struct diff_queue_struct *q = &diff_queued_diff;
3996 int i;
3997 int result = diff_get_patch_id(options, sha1);
3998
3999 for (i = 0; i < q->nr; i++)
4000 diff_free_filepair(q->queue[i]);
4001
4002 free(q->queue);
4003 DIFF_QUEUE_CLEAR(q);
4004
4005 return result;
4006 }
4007
4008 static int is_summary_empty(const struct diff_queue_struct *q)
4009 {
4010 int i;
4011
4012 for (i = 0; i < q->nr; i++) {
4013 const struct diff_filepair *p = q->queue[i];
4014
4015 switch (p->status) {
4016 case DIFF_STATUS_DELETED:
4017 case DIFF_STATUS_ADDED:
4018 case DIFF_STATUS_COPIED:
4019 case DIFF_STATUS_RENAMED:
4020 return 0;
4021 default:
4022 if (p->score)
4023 return 0;
4024 if (p->one->mode && p->two->mode &&
4025 p->one->mode != p->two->mode)
4026 return 0;
4027 break;
4028 }
4029 }
4030 return 1;
4031 }
4032
4033 static const char rename_limit_warning[] =
4034 "inexact rename detection was skipped due to too many files.";
4035
4036 static const char degrade_cc_to_c_warning[] =
4037 "only found copies from modified paths due to too many files.";
4038
4039 static const char rename_limit_advice[] =
4040 "you may want to set your %s variable to at least "
4041 "%d and retry the command.";
4042
4043 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4044 {
4045 if (degraded_cc)
4046 warning(degrade_cc_to_c_warning);
4047 else if (needed)
4048 warning(rename_limit_warning);
4049 else
4050 return;
4051 if (0 < needed && needed < 32767)
4052 warning(rename_limit_advice, varname, needed);
4053 }
4054
4055 void diff_flush(struct diff_options *options)
4056 {
4057 struct diff_queue_struct *q = &diff_queued_diff;
4058 int i, output_format = options->output_format;
4059 int separator = 0;
4060
4061 /*
4062 * Order: raw, stat, summary, patch
4063 * or: name/name-status/checkdiff (other bits clear)
4064 */
4065 if (!q->nr)
4066 goto free_queue;
4067
4068 if (output_format & (DIFF_FORMAT_RAW |
4069 DIFF_FORMAT_NAME |
4070 DIFF_FORMAT_NAME_STATUS |
4071 DIFF_FORMAT_CHECKDIFF)) {
4072 for (i = 0; i < q->nr; i++) {
4073 struct diff_filepair *p = q->queue[i];
4074 if (check_pair_status(p))
4075 flush_one_pair(p, options);
4076 }
4077 separator++;
4078 }
4079
4080 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4081 struct diffstat_t diffstat;
4082
4083 memset(&diffstat, 0, sizeof(struct diffstat_t));
4084 for (i = 0; i < q->nr; i++) {
4085 struct diff_filepair *p = q->queue[i];
4086 if (check_pair_status(p))
4087 diff_flush_stat(p, options, &diffstat);
4088 }
4089 if (output_format & DIFF_FORMAT_NUMSTAT)
4090 show_numstat(&diffstat, options);
4091 if (output_format & DIFF_FORMAT_DIFFSTAT)
4092 show_stats(&diffstat, options);
4093 if (output_format & DIFF_FORMAT_SHORTSTAT)
4094 show_shortstats(&diffstat, options);
4095 free_diffstat_info(&diffstat);
4096 separator++;
4097 }
4098 if (output_format & DIFF_FORMAT_DIRSTAT)
4099 show_dirstat(options);
4100
4101 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4102 for (i = 0; i < q->nr; i++) {
4103 diff_summary(options, q->queue[i]);
4104 }
4105 separator++;
4106 }
4107
4108 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4109 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4110 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4111 /*
4112 * run diff_flush_patch for the exit status. setting
4113 * options->file to /dev/null should be safe, becaue we
4114 * aren't supposed to produce any output anyway.
4115 */
4116 if (options->close_file)
4117 fclose(options->file);
4118 options->file = fopen("/dev/null", "w");
4119 if (!options->file)
4120 die_errno("Could not open /dev/null");
4121 options->close_file = 1;
4122 for (i = 0; i < q->nr; i++) {
4123 struct diff_filepair *p = q->queue[i];
4124 if (check_pair_status(p))
4125 diff_flush_patch(p, options);
4126 if (options->found_changes)
4127 break;
4128 }
4129 }
4130
4131 if (output_format & DIFF_FORMAT_PATCH) {
4132 if (separator) {
4133 putc(options->line_termination, options->file);
4134 if (options->stat_sep) {
4135 /* attach patch instead of inline */
4136 fputs(options->stat_sep, options->file);
4137 }
4138 }
4139
4140 for (i = 0; i < q->nr; i++) {
4141 struct diff_filepair *p = q->queue[i];
4142 if (check_pair_status(p))
4143 diff_flush_patch(p, options);
4144 }
4145 }
4146
4147 if (output_format & DIFF_FORMAT_CALLBACK)
4148 options->format_callback(q, options, options->format_callback_data);
4149
4150 for (i = 0; i < q->nr; i++)
4151 diff_free_filepair(q->queue[i]);
4152 free_queue:
4153 free(q->queue);
4154 DIFF_QUEUE_CLEAR(q);
4155 if (options->close_file)
4156 fclose(options->file);
4157
4158 /*
4159 * Report the content-level differences with HAS_CHANGES;
4160 * diff_addremove/diff_change does not set the bit when
4161 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4162 */
4163 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4164 if (options->found_changes)
4165 DIFF_OPT_SET(options, HAS_CHANGES);
4166 else
4167 DIFF_OPT_CLR(options, HAS_CHANGES);
4168 }
4169 }
4170
4171 static void diffcore_apply_filter(const char *filter)
4172 {
4173 int i;
4174 struct diff_queue_struct *q = &diff_queued_diff;
4175 struct diff_queue_struct outq;
4176 DIFF_QUEUE_CLEAR(&outq);
4177
4178 if (!filter)
4179 return;
4180
4181 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4182 int found;
4183 for (i = found = 0; !found && i < q->nr; i++) {
4184 struct diff_filepair *p = q->queue[i];
4185 if (((p->status == DIFF_STATUS_MODIFIED) &&
4186 ((p->score &&
4187 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4188 (!p->score &&
4189 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4190 ((p->status != DIFF_STATUS_MODIFIED) &&
4191 strchr(filter, p->status)))
4192 found++;
4193 }
4194 if (found)
4195 return;
4196
4197 /* otherwise we will clear the whole queue
4198 * by copying the empty outq at the end of this
4199 * function, but first clear the current entries
4200 * in the queue.
4201 */
4202 for (i = 0; i < q->nr; i++)
4203 diff_free_filepair(q->queue[i]);
4204 }
4205 else {
4206 /* Only the matching ones */
4207 for (i = 0; i < q->nr; i++) {
4208 struct diff_filepair *p = q->queue[i];
4209
4210 if (((p->status == DIFF_STATUS_MODIFIED) &&
4211 ((p->score &&
4212 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4213 (!p->score &&
4214 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4215 ((p->status != DIFF_STATUS_MODIFIED) &&
4216 strchr(filter, p->status)))
4217 diff_q(&outq, p);
4218 else
4219 diff_free_filepair(p);
4220 }
4221 }
4222 free(q->queue);
4223 *q = outq;
4224 }
4225
4226 /* Check whether two filespecs with the same mode and size are identical */
4227 static int diff_filespec_is_identical(struct diff_filespec *one,
4228 struct diff_filespec *two)
4229 {
4230 if (S_ISGITLINK(one->mode))
4231 return 0;
4232 if (diff_populate_filespec(one, 0))
4233 return 0;
4234 if (diff_populate_filespec(two, 0))
4235 return 0;
4236 return !memcmp(one->data, two->data, one->size);
4237 }
4238
4239 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4240 {
4241 int i;
4242 struct diff_queue_struct *q = &diff_queued_diff;
4243 struct diff_queue_struct outq;
4244 DIFF_QUEUE_CLEAR(&outq);
4245
4246 for (i = 0; i < q->nr; i++) {
4247 struct diff_filepair *p = q->queue[i];
4248
4249 /*
4250 * 1. Entries that come from stat info dirtiness
4251 * always have both sides (iow, not create/delete),
4252 * one side of the object name is unknown, with
4253 * the same mode and size. Keep the ones that
4254 * do not match these criteria. They have real
4255 * differences.
4256 *
4257 * 2. At this point, the file is known to be modified,
4258 * with the same mode and size, and the object
4259 * name of one side is unknown. Need to inspect
4260 * the identical contents.
4261 */
4262 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4263 !DIFF_FILE_VALID(p->two) ||
4264 (p->one->sha1_valid && p->two->sha1_valid) ||
4265 (p->one->mode != p->two->mode) ||
4266 diff_populate_filespec(p->one, 1) ||
4267 diff_populate_filespec(p->two, 1) ||
4268 (p->one->size != p->two->size) ||
4269 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4270 diff_q(&outq, p);
4271 else {
4272 /*
4273 * The caller can subtract 1 from skip_stat_unmatch
4274 * to determine how many paths were dirty only
4275 * due to stat info mismatch.
4276 */
4277 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4278 diffopt->skip_stat_unmatch++;
4279 diff_free_filepair(p);
4280 }
4281 }
4282 free(q->queue);
4283 *q = outq;
4284 }
4285
4286 static int diffnamecmp(const void *a_, const void *b_)
4287 {
4288 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4289 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4290 const char *name_a, *name_b;
4291
4292 name_a = a->one ? a->one->path : a->two->path;
4293 name_b = b->one ? b->one->path : b->two->path;
4294 return strcmp(name_a, name_b);
4295 }
4296
4297 void diffcore_fix_diff_index(struct diff_options *options)
4298 {
4299 struct diff_queue_struct *q = &diff_queued_diff;
4300 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4301 }
4302
4303 void diffcore_std(struct diff_options *options)
4304 {
4305 if (options->skip_stat_unmatch)
4306 diffcore_skip_stat_unmatch(options);
4307 if (!options->found_follow) {
4308 /* See try_to_follow_renames() in tree-diff.c */
4309 if (options->break_opt != -1)
4310 diffcore_break(options->break_opt);
4311 if (options->detect_rename)
4312 diffcore_rename(options);
4313 if (options->break_opt != -1)
4314 diffcore_merge_broken();
4315 }
4316 if (options->pickaxe)
4317 diffcore_pickaxe(options);
4318 if (options->orderfile)
4319 diffcore_order(options->orderfile);
4320 if (!options->found_follow)
4321 /* See try_to_follow_renames() in tree-diff.c */
4322 diff_resolve_rename_copy();
4323 diffcore_apply_filter(options->filter);
4324
4325 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4326 DIFF_OPT_SET(options, HAS_CHANGES);
4327 else
4328 DIFF_OPT_CLR(options, HAS_CHANGES);
4329
4330 options->found_follow = 0;
4331 }
4332
4333 int diff_result_code(struct diff_options *opt, int status)
4334 {
4335 int result = 0;
4336
4337 diff_warn_rename_limit("diff.renamelimit",
4338 opt->needed_rename_limit,
4339 opt->degraded_cc_to_c);
4340 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4341 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4342 return status;
4343 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4344 DIFF_OPT_TST(opt, HAS_CHANGES))
4345 result |= 01;
4346 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4347 DIFF_OPT_TST(opt, CHECK_FAILED))
4348 result |= 02;
4349 return result;
4350 }
4351
4352 /*
4353 * Shall changes to this submodule be ignored?
4354 *
4355 * Submodule changes can be configured to be ignored separately for each path,
4356 * but that configuration can be overridden from the command line.
4357 */
4358 static int is_submodule_ignored(const char *path, struct diff_options *options)
4359 {
4360 int ignored = 0;
4361 unsigned orig_flags = options->flags;
4362 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4363 set_diffopt_flags_from_submodule_config(options, path);
4364 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4365 ignored = 1;
4366 options->flags = orig_flags;
4367 return ignored;
4368 }
4369
4370 void diff_addremove(struct diff_options *options,
4371 int addremove, unsigned mode,
4372 const unsigned char *sha1,
4373 const char *concatpath, unsigned dirty_submodule)
4374 {
4375 struct diff_filespec *one, *two;
4376
4377 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4378 return;
4379
4380 /* This may look odd, but it is a preparation for
4381 * feeding "there are unchanged files which should
4382 * not produce diffs, but when you are doing copy
4383 * detection you would need them, so here they are"
4384 * entries to the diff-core. They will be prefixed
4385 * with something like '=' or '*' (I haven't decided
4386 * which but should not make any difference).
4387 * Feeding the same new and old to diff_change()
4388 * also has the same effect.
4389 * Before the final output happens, they are pruned after
4390 * merged into rename/copy pairs as appropriate.
4391 */
4392 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4393 addremove = (addremove == '+' ? '-' :
4394 addremove == '-' ? '+' : addremove);
4395
4396 if (options->prefix &&
4397 strncmp(concatpath, options->prefix, options->prefix_length))
4398 return;
4399
4400 one = alloc_filespec(concatpath);
4401 two = alloc_filespec(concatpath);
4402
4403 if (addremove != '+')
4404 fill_filespec(one, sha1, mode);
4405 if (addremove != '-') {
4406 fill_filespec(two, sha1, mode);
4407 two->dirty_submodule = dirty_submodule;
4408 }
4409
4410 diff_queue(&diff_queued_diff, one, two);
4411 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4412 DIFF_OPT_SET(options, HAS_CHANGES);
4413 }
4414
4415 void diff_change(struct diff_options *options,
4416 unsigned old_mode, unsigned new_mode,
4417 const unsigned char *old_sha1,
4418 const unsigned char *new_sha1,
4419 const char *concatpath,
4420 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4421 {
4422 struct diff_filespec *one, *two;
4423
4424 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4425 is_submodule_ignored(concatpath, options))
4426 return;
4427
4428 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4429 unsigned tmp;
4430 const unsigned char *tmp_c;
4431 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4432 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4433 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4434 new_dirty_submodule = tmp;
4435 }
4436
4437 if (options->prefix &&
4438 strncmp(concatpath, options->prefix, options->prefix_length))
4439 return;
4440
4441 one = alloc_filespec(concatpath);
4442 two = alloc_filespec(concatpath);
4443 fill_filespec(one, old_sha1, old_mode);
4444 fill_filespec(two, new_sha1, new_mode);
4445 one->dirty_submodule = old_dirty_submodule;
4446 two->dirty_submodule = new_dirty_submodule;
4447
4448 diff_queue(&diff_queued_diff, one, two);
4449 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4450 DIFF_OPT_SET(options, HAS_CHANGES);
4451 }
4452
4453 void diff_unmerge(struct diff_options *options,
4454 const char *path,
4455 unsigned mode, const unsigned char *sha1)
4456 {
4457 struct diff_filespec *one, *two;
4458
4459 if (options->prefix &&
4460 strncmp(path, options->prefix, options->prefix_length))
4461 return;
4462
4463 one = alloc_filespec(path);
4464 two = alloc_filespec(path);
4465 fill_filespec(one, sha1, mode);
4466 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4467 }
4468
4469 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4470 size_t *outsize)
4471 {
4472 struct diff_tempfile *temp;
4473 const char *argv[3];
4474 const char **arg = argv;
4475 struct child_process child;
4476 struct strbuf buf = STRBUF_INIT;
4477 int err = 0;
4478
4479 temp = prepare_temp_file(spec->path, spec);
4480 *arg++ = pgm;
4481 *arg++ = temp->name;
4482 *arg = NULL;
4483
4484 memset(&child, 0, sizeof(child));
4485 child.use_shell = 1;
4486 child.argv = argv;
4487 child.out = -1;
4488 if (start_command(&child)) {
4489 remove_tempfile();
4490 return NULL;
4491 }
4492
4493 if (strbuf_read(&buf, child.out, 0) < 0)
4494 err = error("error reading from textconv command '%s'", pgm);
4495 close(child.out);
4496
4497 if (finish_command(&child) || err) {
4498 strbuf_release(&buf);
4499 remove_tempfile();
4500 return NULL;
4501 }
4502 remove_tempfile();
4503
4504 return strbuf_detach(&buf, outsize);
4505 }
4506
4507 size_t fill_textconv(struct userdiff_driver *driver,
4508 struct diff_filespec *df,
4509 char **outbuf)
4510 {
4511 size_t size;
4512
4513 if (!driver || !driver->textconv) {
4514 if (!DIFF_FILE_VALID(df)) {
4515 *outbuf = "";
4516 return 0;
4517 }
4518 if (diff_populate_filespec(df, 0))
4519 die("unable to read files to diff");
4520 *outbuf = df->data;
4521 return df->size;
4522 }
4523
4524 if (driver->textconv_cache && df->sha1_valid) {
4525 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4526 &size);
4527 if (*outbuf)
4528 return size;
4529 }
4530
4531 *outbuf = run_textconv(driver->textconv, df, &size);
4532 if (!*outbuf)
4533 die("unable to read files to diff");
4534
4535 if (driver->textconv_cache && df->sha1_valid) {
4536 /* ignore errors, as we might be in a readonly repository */
4537 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4538 size);
4539 /*
4540 * we could save up changes and flush them all at the end,
4541 * but we would need an extra call after all diffing is done.
4542 * Since generating a cache entry is the slow path anyway,
4543 * this extra overhead probably isn't a big deal.
4544 */
4545 notes_cache_write(driver->textconv_cache);
4546 }
4547
4548 return size;
4549 }
The Best VCS Around