?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Merge branch 'jc/checkout-orphan-warning'
[git/mjg.git] / diff.c
blob66408571abf91d6b90cd69d04164eafd633e407a
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 200;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 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 print_line_count(o->file, lc_a);
585 fprintf(o->file, " +");
586 print_line_count(o->file, lc_b);
587 fprintf(o->file, " @@%s\n", reset);
588 if (lc_a)
589 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
590 if (lc_b)
591 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
592 if (textconv_one)
593 free((char *)data_one);
594 if (textconv_two)
595 free((char *)data_two);
596 }
597
598 struct diff_words_buffer {
599 mmfile_t text;
600 long alloc;
601 struct diff_words_orig {
602 const char *begin, *end;
603 } *orig;
604 int orig_nr, orig_alloc;
605 };
606
607 static void diff_words_append(char *line, unsigned long len,
608 struct diff_words_buffer *buffer)
609 {
610 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
611 line++;
612 len--;
613 memcpy(buffer->text.ptr + buffer->text.size, line, len);
614 buffer->text.size += len;
615 buffer->text.ptr[buffer->text.size] = '\0';
616 }
617
618 struct diff_words_style_elem
619 {
620 const char *prefix;
621 const char *suffix;
622 const char *color; /* NULL; filled in by the setup code if
623 * color is enabled */
624 };
625
626 struct diff_words_style
627 {
628 enum diff_words_type type;
629 struct diff_words_style_elem new, old, ctx;
630 const char *newline;
631 };
632
633 struct diff_words_style diff_words_styles[] = {
634 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
635 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
636 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
637 };
638
639 struct diff_words_data {
640 struct diff_words_buffer minus, plus;
641 const char *current_plus;
642 int last_minus;
643 struct diff_options *opt;
644 regex_t *word_regex;
645 enum diff_words_type type;
646 struct diff_words_style *style;
647 };
648
649 static int fn_out_diff_words_write_helper(FILE *fp,
650 struct diff_words_style_elem *st_el,
651 const char *newline,
652 size_t count, const char *buf,
653 const char *line_prefix)
654 {
655 int print = 0;
656
657 while (count) {
658 char *p = memchr(buf, '\n', count);
659 if (print)
660 fputs(line_prefix, fp);
661 if (p != buf) {
662 if (st_el->color && fputs(st_el->color, fp) < 0)
663 return -1;
664 if (fputs(st_el->prefix, fp) < 0 ||
665 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
666 fputs(st_el->suffix, fp) < 0)
667 return -1;
668 if (st_el->color && *st_el->color
669 && fputs(GIT_COLOR_RESET, fp) < 0)
670 return -1;
671 }
672 if (!p)
673 return 0;
674 if (fputs(newline, fp) < 0)
675 return -1;
676 count -= p + 1 - buf;
677 buf = p + 1;
678 print = 1;
679 }
680 return 0;
681 }
682
683 /*
684 * '--color-words' algorithm can be described as:
685 *
686 * 1. collect a the minus/plus lines of a diff hunk, divided into
687 * minus-lines and plus-lines;
688 *
689 * 2. break both minus-lines and plus-lines into words and
690 * place them into two mmfile_t with one word for each line;
691 *
692 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
693 *
694 * And for the common parts of the both file, we output the plus side text.
695 * diff_words->current_plus is used to trace the current position of the plus file
696 * which printed. diff_words->last_minus is used to trace the last minus word
697 * printed.
698 *
699 * For '--graph' to work with '--color-words', we need to output the graph prefix
700 * on each line of color words output. Generally, there are two conditions on
701 * which we should output the prefix.
702 *
703 * 1. diff_words->last_minus == 0 &&
704 * diff_words->current_plus == diff_words->plus.text.ptr
705 *
706 * that is: the plus text must start as a new line, and if there is no minus
707 * word printed, a graph prefix must be printed.
708 *
709 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
710 * *(diff_words->current_plus - 1) == '\n'
711 *
712 * that is: a graph prefix must be printed following a '\n'
713 */
714 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
715 {
716 if ((diff_words->last_minus == 0 &&
717 diff_words->current_plus == diff_words->plus.text.ptr) ||
718 (diff_words->current_plus > diff_words->plus.text.ptr &&
719 *(diff_words->current_plus - 1) == '\n')) {
720 return 1;
721 } else {
722 return 0;
723 }
724 }
725
726 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
727 {
728 struct diff_words_data *diff_words = priv;
729 struct diff_words_style *style = diff_words->style;
730 int minus_first, minus_len, plus_first, plus_len;
731 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
732 struct diff_options *opt = diff_words->opt;
733 struct strbuf *msgbuf;
734 char *line_prefix = "";
735
736 if (line[0] != '@' || parse_hunk_header(line, len,
737 &minus_first, &minus_len, &plus_first, &plus_len))
738 return;
739
740 assert(opt);
741 if (opt->output_prefix) {
742 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
743 line_prefix = msgbuf->buf;
744 }
745
746 /* POSIX requires that first be decremented by one if len == 0... */
747 if (minus_len) {
748 minus_begin = diff_words->minus.orig[minus_first].begin;
749 minus_end =
750 diff_words->minus.orig[minus_first + minus_len - 1].end;
751 } else
752 minus_begin = minus_end =
753 diff_words->minus.orig[minus_first].end;
754
755 if (plus_len) {
756 plus_begin = diff_words->plus.orig[plus_first].begin;
757 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
758 } else
759 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
760
761 if (color_words_output_graph_prefix(diff_words)) {
762 fputs(line_prefix, diff_words->opt->file);
763 }
764 if (diff_words->current_plus != plus_begin) {
765 fn_out_diff_words_write_helper(diff_words->opt->file,
766 &style->ctx, style->newline,
767 plus_begin - diff_words->current_plus,
768 diff_words->current_plus, line_prefix);
769 if (*(plus_begin - 1) == '\n')
770 fputs(line_prefix, diff_words->opt->file);
771 }
772 if (minus_begin != minus_end) {
773 fn_out_diff_words_write_helper(diff_words->opt->file,
774 &style->old, style->newline,
775 minus_end - minus_begin, minus_begin,
776 line_prefix);
777 }
778 if (plus_begin != plus_end) {
779 fn_out_diff_words_write_helper(diff_words->opt->file,
780 &style->new, style->newline,
781 plus_end - plus_begin, plus_begin,
782 line_prefix);
783 }
784
785 diff_words->current_plus = plus_end;
786 diff_words->last_minus = minus_first;
787 }
788
789 /* This function starts looking at *begin, and returns 0 iff a word was found. */
790 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
791 int *begin, int *end)
792 {
793 if (word_regex && *begin < buffer->size) {
794 regmatch_t match[1];
795 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
796 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
797 '\n', match[0].rm_eo - match[0].rm_so);
798 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
799 *begin += match[0].rm_so;
800 return *begin >= *end;
801 }
802 return -1;
803 }
804
805 /* find the next word */
806 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
807 (*begin)++;
808 if (*begin >= buffer->size)
809 return -1;
810
811 /* find the end of the word */
812 *end = *begin + 1;
813 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
814 (*end)++;
815
816 return 0;
817 }
818
819 /*
820 * This function splits the words in buffer->text, stores the list with
821 * newline separator into out, and saves the offsets of the original words
822 * in buffer->orig.
823 */
824 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
825 regex_t *word_regex)
826 {
827 int i, j;
828 long alloc = 0;
829
830 out->size = 0;
831 out->ptr = NULL;
832
833 /* fake an empty "0th" word */
834 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
835 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
836 buffer->orig_nr = 1;
837
838 for (i = 0; i < buffer->text.size; i++) {
839 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
840 return;
841
842 /* store original boundaries */
843 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
844 buffer->orig_alloc);
845 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
846 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
847 buffer->orig_nr++;
848
849 /* store one word */
850 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
851 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
852 out->ptr[out->size + j - i] = '\n';
853 out->size += j - i + 1;
854
855 i = j - 1;
856 }
857 }
858
859 /* this executes the word diff on the accumulated buffers */
860 static void diff_words_show(struct diff_words_data *diff_words)
861 {
862 xpparam_t xpp;
863 xdemitconf_t xecfg;
864 mmfile_t minus, plus;
865 struct diff_words_style *style = diff_words->style;
866
867 struct diff_options *opt = diff_words->opt;
868 struct strbuf *msgbuf;
869 char *line_prefix = "";
870
871 assert(opt);
872 if (opt->output_prefix) {
873 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
874 line_prefix = msgbuf->buf;
875 }
876
877 /* special case: only removal */
878 if (!diff_words->plus.text.size) {
879 fputs(line_prefix, diff_words->opt->file);
880 fn_out_diff_words_write_helper(diff_words->opt->file,
881 &style->old, style->newline,
882 diff_words->minus.text.size,
883 diff_words->minus.text.ptr, line_prefix);
884 diff_words->minus.text.size = 0;
885 return;
886 }
887
888 diff_words->current_plus = diff_words->plus.text.ptr;
889 diff_words->last_minus = 0;
890
891 memset(&xpp, 0, sizeof(xpp));
892 memset(&xecfg, 0, sizeof(xecfg));
893 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
894 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
895 xpp.flags = 0;
896 /* as only the hunk header will be parsed, we need a 0-context */
897 xecfg.ctxlen = 0;
898 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
899 &xpp, &xecfg);
900 free(minus.ptr);
901 free(plus.ptr);
902 if (diff_words->current_plus != diff_words->plus.text.ptr +
903 diff_words->plus.text.size) {
904 if (color_words_output_graph_prefix(diff_words))
905 fputs(line_prefix, diff_words->opt->file);
906 fn_out_diff_words_write_helper(diff_words->opt->file,
907 &style->ctx, style->newline,
908 diff_words->plus.text.ptr + diff_words->plus.text.size
909 - diff_words->current_plus, diff_words->current_plus,
910 line_prefix);
911 }
912 diff_words->minus.text.size = diff_words->plus.text.size = 0;
913 }
914
915 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
916 static void diff_words_flush(struct emit_callback *ecbdata)
917 {
918 if (ecbdata->diff_words->minus.text.size ||
919 ecbdata->diff_words->plus.text.size)
920 diff_words_show(ecbdata->diff_words);
921 }
922
923 static void free_diff_words_data(struct emit_callback *ecbdata)
924 {
925 if (ecbdata->diff_words) {
926 diff_words_flush(ecbdata);
927 free (ecbdata->diff_words->minus.text.ptr);
928 free (ecbdata->diff_words->minus.orig);
929 free (ecbdata->diff_words->plus.text.ptr);
930 free (ecbdata->diff_words->plus.orig);
931 if (ecbdata->diff_words->word_regex) {
932 regfree(ecbdata->diff_words->word_regex);
933 free(ecbdata->diff_words->word_regex);
934 }
935 free(ecbdata->diff_words);
936 ecbdata->diff_words = NULL;
937 }
938 }
939
940 const char *diff_get_color(int diff_use_color, enum color_diff ix)
941 {
942 if (diff_use_color)
943 return diff_colors[ix];
944 return "";
945 }
946
947 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
948 {
949 const char *cp;
950 unsigned long allot;
951 size_t l = len;
952
953 if (ecb->truncate)
954 return ecb->truncate(line, len);
955 cp = line;
956 allot = l;
957 while (0 < l) {
958 (void) utf8_width(&cp, &l);
959 if (!cp)
960 break; /* truncated in the middle? */
961 }
962 return allot - l;
963 }
964
965 static void find_lno(const char *line, struct emit_callback *ecbdata)
966 {
967 const char *p;
968 ecbdata->lno_in_preimage = 0;
969 ecbdata->lno_in_postimage = 0;
970 p = strchr(line, '-');
971 if (!p)
972 return; /* cannot happen */
973 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
974 p = strchr(p, '+');
975 if (!p)
976 return; /* cannot happen */
977 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
978 }
979
980 static void fn_out_consume(void *priv, char *line, unsigned long len)
981 {
982 struct emit_callback *ecbdata = priv;
983 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
984 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
985 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
986 struct diff_options *o = ecbdata->opt;
987 char *line_prefix = "";
988 struct strbuf *msgbuf;
989
990 if (o && o->output_prefix) {
991 msgbuf = o->output_prefix(o, o->output_prefix_data);
992 line_prefix = msgbuf->buf;
993 }
994
995 if (ecbdata->header) {
996 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
997 strbuf_reset(ecbdata->header);
998 ecbdata->header = NULL;
999 }
1000 *(ecbdata->found_changesp) = 1;
1001
1002 if (ecbdata->label_path[0]) {
1003 const char *name_a_tab, *name_b_tab;
1004
1005 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1006 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1007
1008 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1009 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1010 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1011 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1012 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1013 }
1014
1015 if (diff_suppress_blank_empty
1016 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1017 line[0] = '\n';
1018 len = 1;
1019 }
1020
1021 if (line[0] == '@') {
1022 if (ecbdata->diff_words)
1023 diff_words_flush(ecbdata);
1024 len = sane_truncate_line(ecbdata, line, len);
1025 find_lno(line, ecbdata);
1026 emit_hunk_header(ecbdata, line, len);
1027 if (line[len-1] != '\n')
1028 putc('\n', ecbdata->opt->file);
1029 return;
1030 }
1031
1032 if (len < 1) {
1033 emit_line(ecbdata->opt, reset, reset, line, len);
1034 if (ecbdata->diff_words
1035 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1036 fputs("~\n", ecbdata->opt->file);
1037 return;
1038 }
1039
1040 if (ecbdata->diff_words) {
1041 if (line[0] == '-') {
1042 diff_words_append(line, len,
1043 &ecbdata->diff_words->minus);
1044 return;
1045 } else if (line[0] == '+') {
1046 diff_words_append(line, len,
1047 &ecbdata->diff_words->plus);
1048 return;
1049 }
1050 diff_words_flush(ecbdata);
1051 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1052 emit_line(ecbdata->opt, plain, reset, line, len);
1053 fputs("~\n", ecbdata->opt->file);
1054 } else {
1055 /* don't print the prefix character */
1056 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1057 }
1058 return;
1059 }
1060
1061 if (line[0] != '+') {
1062 const char *color =
1063 diff_get_color(ecbdata->color_diff,
1064 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1065 ecbdata->lno_in_preimage++;
1066 if (line[0] == ' ')
1067 ecbdata->lno_in_postimage++;
1068 emit_line(ecbdata->opt, color, reset, line, len);
1069 } else {
1070 ecbdata->lno_in_postimage++;
1071 emit_add_line(reset, ecbdata, line + 1, len - 1);
1072 }
1073 }
1074
1075 static char *pprint_rename(const char *a, const char *b)
1076 {
1077 const char *old = a;
1078 const char *new = b;
1079 struct strbuf name = STRBUF_INIT;
1080 int pfx_length, sfx_length;
1081 int len_a = strlen(a);
1082 int len_b = strlen(b);
1083 int a_midlen, b_midlen;
1084 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1085 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1086
1087 if (qlen_a || qlen_b) {
1088 quote_c_style(a, &name, NULL, 0);
1089 strbuf_addstr(&name, " => ");
1090 quote_c_style(b, &name, NULL, 0);
1091 return strbuf_detach(&name, NULL);
1092 }
1093
1094 /* Find common prefix */
1095 pfx_length = 0;
1096 while (*old && *new && *old == *new) {
1097 if (*old == '/')
1098 pfx_length = old - a + 1;
1099 old++;
1100 new++;
1101 }
1102
1103 /* Find common suffix */
1104 old = a + len_a;
1105 new = b + len_b;
1106 sfx_length = 0;
1107 while (a <= old && b <= new && *old == *new) {
1108 if (*old == '/')
1109 sfx_length = len_a - (old - a);
1110 old--;
1111 new--;
1112 }
1113
1114 /*
1115 * pfx{mid-a => mid-b}sfx
1116 * {pfx-a => pfx-b}sfx
1117 * pfx{sfx-a => sfx-b}
1118 * name-a => name-b
1119 */
1120 a_midlen = len_a - pfx_length - sfx_length;
1121 b_midlen = len_b - pfx_length - sfx_length;
1122 if (a_midlen < 0)
1123 a_midlen = 0;
1124 if (b_midlen < 0)
1125 b_midlen = 0;
1126
1127 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1128 if (pfx_length + sfx_length) {
1129 strbuf_add(&name, a, pfx_length);
1130 strbuf_addch(&name, '{');
1131 }
1132 strbuf_add(&name, a + pfx_length, a_midlen);
1133 strbuf_addstr(&name, " => ");
1134 strbuf_add(&name, b + pfx_length, b_midlen);
1135 if (pfx_length + sfx_length) {
1136 strbuf_addch(&name, '}');
1137 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1138 }
1139 return strbuf_detach(&name, NULL);
1140 }
1141
1142 struct diffstat_t {
1143 int nr;
1144 int alloc;
1145 struct diffstat_file {
1146 char *from_name;
1147 char *name;
1148 char *print_name;
1149 unsigned is_unmerged:1;
1150 unsigned is_binary:1;
1151 unsigned is_renamed:1;
1152 uintmax_t added, deleted;
1153 } **files;
1154 };
1155
1156 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1157 const char *name_a,
1158 const char *name_b)
1159 {
1160 struct diffstat_file *x;
1161 x = xcalloc(sizeof (*x), 1);
1162 if (diffstat->nr == diffstat->alloc) {
1163 diffstat->alloc = alloc_nr(diffstat->alloc);
1164 diffstat->files = xrealloc(diffstat->files,
1165 diffstat->alloc * sizeof(x));
1166 }
1167 diffstat->files[diffstat->nr++] = x;
1168 if (name_b) {
1169 x->from_name = xstrdup(name_a);
1170 x->name = xstrdup(name_b);
1171 x->is_renamed = 1;
1172 }
1173 else {
1174 x->from_name = NULL;
1175 x->name = xstrdup(name_a);
1176 }
1177 return x;
1178 }
1179
1180 static void diffstat_consume(void *priv, char *line, unsigned long len)
1181 {
1182 struct diffstat_t *diffstat = priv;
1183 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1184
1185 if (line[0] == '+')
1186 x->added++;
1187 else if (line[0] == '-')
1188 x->deleted++;
1189 }
1190
1191 const char mime_boundary_leader[] = "------------";
1192
1193 static int scale_linear(int it, int width, int max_change)
1194 {
1195 /*
1196 * make sure that at least one '-' is printed if there were deletions,
1197 * and likewise for '+'.
1198 */
1199 if (max_change < 2)
1200 return it;
1201 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1202 }
1203
1204 static void show_name(FILE *file,
1205 const char *prefix, const char *name, int len)
1206 {
1207 fprintf(file, " %s%-*s |", prefix, len, name);
1208 }
1209
1210 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1211 {
1212 if (cnt <= 0)
1213 return;
1214 fprintf(file, "%s", set);
1215 while (cnt--)
1216 putc(ch, file);
1217 fprintf(file, "%s", reset);
1218 }
1219
1220 static void fill_print_name(struct diffstat_file *file)
1221 {
1222 char *pname;
1223
1224 if (file->print_name)
1225 return;
1226
1227 if (!file->is_renamed) {
1228 struct strbuf buf = STRBUF_INIT;
1229 if (quote_c_style(file->name, &buf, NULL, 0)) {
1230 pname = strbuf_detach(&buf, NULL);
1231 } else {
1232 pname = file->name;
1233 strbuf_release(&buf);
1234 }
1235 } else {
1236 pname = pprint_rename(file->from_name, file->name);
1237 }
1238 file->print_name = pname;
1239 }
1240
1241 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1242 {
1243 int i, len, add, del, adds = 0, dels = 0;
1244 uintmax_t max_change = 0, max_len = 0;
1245 int total_files = data->nr;
1246 int width, name_width;
1247 const char *reset, *set, *add_c, *del_c;
1248 const char *line_prefix = "";
1249 struct strbuf *msg = NULL;
1250
1251 if (data->nr == 0)
1252 return;
1253
1254 if (options->output_prefix) {
1255 msg = options->output_prefix(options, options->output_prefix_data);
1256 line_prefix = msg->buf;
1257 }
1258
1259 width = options->stat_width ? options->stat_width : 80;
1260 name_width = options->stat_name_width ? options->stat_name_width : 50;
1261
1262 /* Sanity: give at least 5 columns to the graph,
1263 * but leave at least 10 columns for the name.
1264 */
1265 if (width < 25)
1266 width = 25;
1267 if (name_width < 10)
1268 name_width = 10;
1269 else if (width < name_width + 15)
1270 name_width = width - 15;
1271
1272 /* Find the longest filename and max number of changes */
1273 reset = diff_get_color_opt(options, DIFF_RESET);
1274 set = diff_get_color_opt(options, DIFF_PLAIN);
1275 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1276 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1277
1278 for (i = 0; i < data->nr; i++) {
1279 struct diffstat_file *file = data->files[i];
1280 uintmax_t change = file->added + file->deleted;
1281 fill_print_name(file);
1282 len = strlen(file->print_name);
1283 if (max_len < len)
1284 max_len = len;
1285
1286 if (file->is_binary || file->is_unmerged)
1287 continue;
1288 if (max_change < change)
1289 max_change = change;
1290 }
1291
1292 /* Compute the width of the graph part;
1293 * 10 is for one blank at the beginning of the line plus
1294 * " | count " between the name and the graph.
1295 *
1296 * From here on, name_width is the width of the name area,
1297 * and width is the width of the graph area.
1298 */
1299 name_width = (name_width < max_len) ? name_width : max_len;
1300 if (width < (name_width + 10) + max_change)
1301 width = width - (name_width + 10);
1302 else
1303 width = max_change;
1304
1305 for (i = 0; i < data->nr; i++) {
1306 const char *prefix = "";
1307 char *name = data->files[i]->print_name;
1308 uintmax_t added = data->files[i]->added;
1309 uintmax_t deleted = data->files[i]->deleted;
1310 int name_len;
1311
1312 /*
1313 * "scale" the filename
1314 */
1315 len = name_width;
1316 name_len = strlen(name);
1317 if (name_width < name_len) {
1318 char *slash;
1319 prefix = "...";
1320 len -= 3;
1321 name += name_len - len;
1322 slash = strchr(name, '/');
1323 if (slash)
1324 name = slash;
1325 }
1326
1327 if (data->files[i]->is_binary) {
1328 fprintf(options->file, "%s", line_prefix);
1329 show_name(options->file, prefix, name, len);
1330 fprintf(options->file, " Bin ");
1331 fprintf(options->file, "%s%"PRIuMAX"%s",
1332 del_c, deleted, reset);
1333 fprintf(options->file, " -> ");
1334 fprintf(options->file, "%s%"PRIuMAX"%s",
1335 add_c, added, reset);
1336 fprintf(options->file, " bytes");
1337 fprintf(options->file, "\n");
1338 continue;
1339 }
1340 else if (data->files[i]->is_unmerged) {
1341 fprintf(options->file, "%s", line_prefix);
1342 show_name(options->file, prefix, name, len);
1343 fprintf(options->file, " Unmerged\n");
1344 continue;
1345 }
1346 else if (!data->files[i]->is_renamed &&
1347 (added + deleted == 0)) {
1348 total_files--;
1349 continue;
1350 }
1351
1352 /*
1353 * scale the add/delete
1354 */
1355 add = added;
1356 del = deleted;
1357 adds += add;
1358 dels += del;
1359
1360 if (width <= max_change) {
1361 add = scale_linear(add, width, max_change);
1362 del = scale_linear(del, width, max_change);
1363 }
1364 fprintf(options->file, "%s", line_prefix);
1365 show_name(options->file, prefix, name, len);
1366 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1367 added + deleted ? " " : "");
1368 show_graph(options->file, '+', add, add_c, reset);
1369 show_graph(options->file, '-', del, del_c, reset);
1370 fprintf(options->file, "\n");
1371 }
1372 fprintf(options->file, "%s", line_prefix);
1373 fprintf(options->file,
1374 " %d files changed, %d insertions(+), %d deletions(-)\n",
1375 total_files, adds, dels);
1376 }
1377
1378 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1379 {
1380 int i, adds = 0, dels = 0, total_files = data->nr;
1381
1382 if (data->nr == 0)
1383 return;
1384
1385 for (i = 0; i < data->nr; i++) {
1386 if (!data->files[i]->is_binary &&
1387 !data->files[i]->is_unmerged) {
1388 int added = data->files[i]->added;
1389 int deleted= data->files[i]->deleted;
1390 if (!data->files[i]->is_renamed &&
1391 (added + deleted == 0)) {
1392 total_files--;
1393 } else {
1394 adds += added;
1395 dels += deleted;
1396 }
1397 }
1398 }
1399 if (options->output_prefix) {
1400 struct strbuf *msg = NULL;
1401 msg = options->output_prefix(options,
1402 options->output_prefix_data);
1403 fprintf(options->file, "%s", msg->buf);
1404 }
1405 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1406 total_files, adds, dels);
1407 }
1408
1409 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1410 {
1411 int i;
1412
1413 if (data->nr == 0)
1414 return;
1415
1416 for (i = 0; i < data->nr; i++) {
1417 struct diffstat_file *file = data->files[i];
1418
1419 if (options->output_prefix) {
1420 struct strbuf *msg = NULL;
1421 msg = options->output_prefix(options,
1422 options->output_prefix_data);
1423 fprintf(options->file, "%s", msg->buf);
1424 }
1425
1426 if (file->is_binary)
1427 fprintf(options->file, "-\t-\t");
1428 else
1429 fprintf(options->file,
1430 "%"PRIuMAX"\t%"PRIuMAX"\t",
1431 file->added, file->deleted);
1432 if (options->line_termination) {
1433 fill_print_name(file);
1434 if (!file->is_renamed)
1435 write_name_quoted(file->name, options->file,
1436 options->line_termination);
1437 else {
1438 fputs(file->print_name, options->file);
1439 putc(options->line_termination, options->file);
1440 }
1441 } else {
1442 if (file->is_renamed) {
1443 putc('\0', options->file);
1444 write_name_quoted(file->from_name, options->file, '\0');
1445 }
1446 write_name_quoted(file->name, options->file, '\0');
1447 }
1448 }
1449 }
1450
1451 struct dirstat_file {
1452 const char *name;
1453 unsigned long changed;
1454 };
1455
1456 struct dirstat_dir {
1457 struct dirstat_file *files;
1458 int alloc, nr, percent, cumulative;
1459 };
1460
1461 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1462 unsigned long changed, const char *base, int baselen)
1463 {
1464 unsigned long this_dir = 0;
1465 unsigned int sources = 0;
1466 const char *line_prefix = "";
1467 struct strbuf *msg = NULL;
1468
1469 if (opt->output_prefix) {
1470 msg = opt->output_prefix(opt, opt->output_prefix_data);
1471 line_prefix = msg->buf;
1472 }
1473
1474 while (dir->nr) {
1475 struct dirstat_file *f = dir->files;
1476 int namelen = strlen(f->name);
1477 unsigned long this;
1478 char *slash;
1479
1480 if (namelen < baselen)
1481 break;
1482 if (memcmp(f->name, base, baselen))
1483 break;
1484 slash = strchr(f->name + baselen, '/');
1485 if (slash) {
1486 int newbaselen = slash + 1 - f->name;
1487 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1488 sources++;
1489 } else {
1490 this = f->changed;
1491 dir->files++;
1492 dir->nr--;
1493 sources += 2;
1494 }
1495 this_dir += this;
1496 }
1497
1498 /*
1499 * We don't report dirstat's for
1500 * - the top level
1501 * - or cases where everything came from a single directory
1502 * under this directory (sources == 1).
1503 */
1504 if (baselen && sources != 1) {
1505 int permille = this_dir * 1000 / changed;
1506 if (permille) {
1507 int percent = permille / 10;
1508 if (percent >= dir->percent) {
1509 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1510 percent, permille % 10, baselen, base);
1511 if (!dir->cumulative)
1512 return 0;
1513 }
1514 }
1515 }
1516 return this_dir;
1517 }
1518
1519 static int dirstat_compare(const void *_a, const void *_b)
1520 {
1521 const struct dirstat_file *a = _a;
1522 const struct dirstat_file *b = _b;
1523 return strcmp(a->name, b->name);
1524 }
1525
1526 static void show_dirstat(struct diff_options *options)
1527 {
1528 int i;
1529 unsigned long changed;
1530 struct dirstat_dir dir;
1531 struct diff_queue_struct *q = &diff_queued_diff;
1532
1533 dir.files = NULL;
1534 dir.alloc = 0;
1535 dir.nr = 0;
1536 dir.percent = options->dirstat_percent;
1537 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1538
1539 changed = 0;
1540 for (i = 0; i < q->nr; i++) {
1541 struct diff_filepair *p = q->queue[i];
1542 const char *name;
1543 unsigned long copied, added, damage;
1544
1545 name = p->one->path ? p->one->path : p->two->path;
1546
1547 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1548 diff_populate_filespec(p->one, 0);
1549 diff_populate_filespec(p->two, 0);
1550 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1551 &copied, &added);
1552 diff_free_filespec_data(p->one);
1553 diff_free_filespec_data(p->two);
1554 } else if (DIFF_FILE_VALID(p->one)) {
1555 diff_populate_filespec(p->one, 1);
1556 copied = added = 0;
1557 diff_free_filespec_data(p->one);
1558 } else if (DIFF_FILE_VALID(p->two)) {
1559 diff_populate_filespec(p->two, 1);
1560 copied = 0;
1561 added = p->two->size;
1562 diff_free_filespec_data(p->two);
1563 } else
1564 continue;
1565
1566 /*
1567 * Original minus copied is the removed material,
1568 * added is the new material. They are both damages
1569 * made to the preimage. In --dirstat-by-file mode, count
1570 * damaged files, not damaged lines. This is done by
1571 * counting only a single damaged line per file.
1572 */
1573 damage = (p->one->size - copied) + added;
1574 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE) && damage > 0)
1575 damage = 1;
1576
1577 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1578 dir.files[dir.nr].name = name;
1579 dir.files[dir.nr].changed = damage;
1580 changed += damage;
1581 dir.nr++;
1582 }
1583
1584 /* This can happen even with many files, if everything was renames */
1585 if (!changed)
1586 return;
1587
1588 /* Show all directories with more than x% of the changes */
1589 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1590 gather_dirstat(options, &dir, changed, "", 0);
1591 }
1592
1593 static void free_diffstat_info(struct diffstat_t *diffstat)
1594 {
1595 int i;
1596 for (i = 0; i < diffstat->nr; i++) {
1597 struct diffstat_file *f = diffstat->files[i];
1598 if (f->name != f->print_name)
1599 free(f->print_name);
1600 free(f->name);
1601 free(f->from_name);
1602 free(f);
1603 }
1604 free(diffstat->files);
1605 }
1606
1607 struct checkdiff_t {
1608 const char *filename;
1609 int lineno;
1610 int conflict_marker_size;
1611 struct diff_options *o;
1612 unsigned ws_rule;
1613 unsigned status;
1614 };
1615
1616 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1617 {
1618 char firstchar;
1619 int cnt;
1620
1621 if (len < marker_size + 1)
1622 return 0;
1623 firstchar = line[0];
1624 switch (firstchar) {
1625 case '=': case '>': case '<': case '|':
1626 break;
1627 default:
1628 return 0;
1629 }
1630 for (cnt = 1; cnt < marker_size; cnt++)
1631 if (line[cnt] != firstchar)
1632 return 0;
1633 /* line[1] thru line[marker_size-1] are same as firstchar */
1634 if (len < marker_size + 1 || !isspace(line[marker_size]))
1635 return 0;
1636 return 1;
1637 }
1638
1639 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1640 {
1641 struct checkdiff_t *data = priv;
1642 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1643 int marker_size = data->conflict_marker_size;
1644 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1645 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1646 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1647 char *err;
1648 char *line_prefix = "";
1649 struct strbuf *msgbuf;
1650
1651 assert(data->o);
1652 if (data->o->output_prefix) {
1653 msgbuf = data->o->output_prefix(data->o,
1654 data->o->output_prefix_data);
1655 line_prefix = msgbuf->buf;
1656 }
1657
1658 if (line[0] == '+') {
1659 unsigned bad;
1660 data->lineno++;
1661 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1662 data->status |= 1;
1663 fprintf(data->o->file,
1664 "%s%s:%d: leftover conflict marker\n",
1665 line_prefix, data->filename, data->lineno);
1666 }
1667 bad = ws_check(line + 1, len - 1, data->ws_rule);
1668 if (!bad)
1669 return;
1670 data->status |= bad;
1671 err = whitespace_error_string(bad);
1672 fprintf(data->o->file, "%s%s:%d: %s.\n",
1673 line_prefix, data->filename, data->lineno, err);
1674 free(err);
1675 emit_line(data->o, set, reset, line, 1);
1676 ws_check_emit(line + 1, len - 1, data->ws_rule,
1677 data->o->file, set, reset, ws);
1678 } else if (line[0] == ' ') {
1679 data->lineno++;
1680 } else if (line[0] == '@') {
1681 char *plus = strchr(line, '+');
1682 if (plus)
1683 data->lineno = strtol(plus, NULL, 10) - 1;
1684 else
1685 die("invalid diff");
1686 }
1687 }
1688
1689 static unsigned char *deflate_it(char *data,
1690 unsigned long size,
1691 unsigned long *result_size)
1692 {
1693 int bound;
1694 unsigned char *deflated;
1695 z_stream stream;
1696
1697 memset(&stream, 0, sizeof(stream));
1698 deflateInit(&stream, zlib_compression_level);
1699 bound = deflateBound(&stream, size);
1700 deflated = xmalloc(bound);
1701 stream.next_out = deflated;
1702 stream.avail_out = bound;
1703
1704 stream.next_in = (unsigned char *)data;
1705 stream.avail_in = size;
1706 while (deflate(&stream, Z_FINISH) == Z_OK)
1707 ; /* nothing */
1708 deflateEnd(&stream);
1709 *result_size = stream.total_out;
1710 return deflated;
1711 }
1712
1713 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1714 {
1715 void *cp;
1716 void *delta;
1717 void *deflated;
1718 void *data;
1719 unsigned long orig_size;
1720 unsigned long delta_size;
1721 unsigned long deflate_size;
1722 unsigned long data_size;
1723
1724 /* We could do deflated delta, or we could do just deflated two,
1725 * whichever is smaller.
1726 */
1727 delta = NULL;
1728 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1729 if (one->size && two->size) {
1730 delta = diff_delta(one->ptr, one->size,
1731 two->ptr, two->size,
1732 &delta_size, deflate_size);
1733 if (delta) {
1734 void *to_free = delta;
1735 orig_size = delta_size;
1736 delta = deflate_it(delta, delta_size, &delta_size);
1737 free(to_free);
1738 }
1739 }
1740
1741 if (delta && delta_size < deflate_size) {
1742 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1743 free(deflated);
1744 data = delta;
1745 data_size = delta_size;
1746 }
1747 else {
1748 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1749 free(delta);
1750 data = deflated;
1751 data_size = deflate_size;
1752 }
1753
1754 /* emit data encoded in base85 */
1755 cp = data;
1756 while (data_size) {
1757 int bytes = (52 < data_size) ? 52 : data_size;
1758 char line[70];
1759 data_size -= bytes;
1760 if (bytes <= 26)
1761 line[0] = bytes + 'A' - 1;
1762 else
1763 line[0] = bytes - 26 + 'a' - 1;
1764 encode_85(line + 1, cp, bytes);
1765 cp = (char *) cp + bytes;
1766 fprintf(file, "%s", prefix);
1767 fputs(line, file);
1768 fputc('\n', file);
1769 }
1770 fprintf(file, "%s\n", prefix);
1771 free(data);
1772 }
1773
1774 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1775 {
1776 fprintf(file, "%sGIT binary patch\n", prefix);
1777 emit_binary_diff_body(file, one, two, prefix);
1778 emit_binary_diff_body(file, two, one, prefix);
1779 }
1780
1781 static void diff_filespec_load_driver(struct diff_filespec *one)
1782 {
1783 /* Use already-loaded driver */
1784 if (one->driver)
1785 return;
1786
1787 if (S_ISREG(one->mode))
1788 one->driver = userdiff_find_by_path(one->path);
1789
1790 /* Fallback to default settings */
1791 if (!one->driver)
1792 one->driver = userdiff_find_by_name("default");
1793 }
1794
1795 int diff_filespec_is_binary(struct diff_filespec *one)
1796 {
1797 if (one->is_binary == -1) {
1798 diff_filespec_load_driver(one);
1799 if (one->driver->binary != -1)
1800 one->is_binary = one->driver->binary;
1801 else {
1802 if (!one->data && DIFF_FILE_VALID(one))
1803 diff_populate_filespec(one, 0);
1804 if (one->data)
1805 one->is_binary = buffer_is_binary(one->data,
1806 one->size);
1807 if (one->is_binary == -1)
1808 one->is_binary = 0;
1809 }
1810 }
1811 return one->is_binary;
1812 }
1813
1814 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1815 {
1816 diff_filespec_load_driver(one);
1817 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1818 }
1819
1820 static const char *userdiff_word_regex(struct diff_filespec *one)
1821 {
1822 diff_filespec_load_driver(one);
1823 return one->driver->word_regex;
1824 }
1825
1826 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1827 {
1828 if (!options->a_prefix)
1829 options->a_prefix = a;
1830 if (!options->b_prefix)
1831 options->b_prefix = b;
1832 }
1833
1834 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1835 {
1836 if (!DIFF_FILE_VALID(one))
1837 return NULL;
1838
1839 diff_filespec_load_driver(one);
1840 if (!one->driver->textconv)
1841 return NULL;
1842
1843 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1844 struct notes_cache *c = xmalloc(sizeof(*c));
1845 struct strbuf name = STRBUF_INIT;
1846
1847 strbuf_addf(&name, "textconv/%s", one->driver->name);
1848 notes_cache_init(c, name.buf, one->driver->textconv);
1849 one->driver->textconv_cache = c;
1850 }
1851
1852 return one->driver;
1853 }
1854
1855 static void builtin_diff(const char *name_a,
1856 const char *name_b,
1857 struct diff_filespec *one,
1858 struct diff_filespec *two,
1859 const char *xfrm_msg,
1860 int must_show_header,
1861 struct diff_options *o,
1862 int complete_rewrite)
1863 {
1864 mmfile_t mf1, mf2;
1865 const char *lbl[2];
1866 char *a_one, *b_two;
1867 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1868 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1869 const char *a_prefix, *b_prefix;
1870 struct userdiff_driver *textconv_one = NULL;
1871 struct userdiff_driver *textconv_two = NULL;
1872 struct strbuf header = STRBUF_INIT;
1873 struct strbuf *msgbuf;
1874 char *line_prefix = "";
1875
1876 if (o->output_prefix) {
1877 msgbuf = o->output_prefix(o, o->output_prefix_data);
1878 line_prefix = msgbuf->buf;
1879 }
1880
1881 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1882 (!one->mode || S_ISGITLINK(one->mode)) &&
1883 (!two->mode || S_ISGITLINK(two->mode))) {
1884 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1885 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1886 show_submodule_summary(o->file, one ? one->path : two->path,
1887 one->sha1, two->sha1, two->dirty_submodule,
1888 del, add, reset);
1889 return;
1890 }
1891
1892 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1893 textconv_one = get_textconv(one);
1894 textconv_two = get_textconv(two);
1895 }
1896
1897 diff_set_mnemonic_prefix(o, "a/", "b/");
1898 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1899 a_prefix = o->b_prefix;
1900 b_prefix = o->a_prefix;
1901 } else {
1902 a_prefix = o->a_prefix;
1903 b_prefix = o->b_prefix;
1904 }
1905
1906 /* Never use a non-valid filename anywhere if at all possible */
1907 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1908 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1909
1910 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1911 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1912 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1913 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1914 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1915 if (lbl[0][0] == '/') {
1916 /* /dev/null */
1917 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1918 if (xfrm_msg)
1919 strbuf_addstr(&header, xfrm_msg);
1920 must_show_header = 1;
1921 }
1922 else if (lbl[1][0] == '/') {
1923 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1924 if (xfrm_msg)
1925 strbuf_addstr(&header, xfrm_msg);
1926 must_show_header = 1;
1927 }
1928 else {
1929 if (one->mode != two->mode) {
1930 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1931 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1932 must_show_header = 1;
1933 }
1934 if (xfrm_msg)
1935 strbuf_addstr(&header, xfrm_msg);
1936
1937 /*
1938 * we do not run diff between different kind
1939 * of objects.
1940 */
1941 if ((one->mode ^ two->mode) & S_IFMT)
1942 goto free_ab_and_return;
1943 if (complete_rewrite &&
1944 (textconv_one || !diff_filespec_is_binary(one)) &&
1945 (textconv_two || !diff_filespec_is_binary(two))) {
1946 fprintf(o->file, "%s", header.buf);
1947 strbuf_reset(&header);
1948 emit_rewrite_diff(name_a, name_b, one, two,
1949 textconv_one, textconv_two, o);
1950 o->found_changes = 1;
1951 goto free_ab_and_return;
1952 }
1953 }
1954
1955 if (!DIFF_OPT_TST(o, TEXT) &&
1956 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1957 (!textconv_two && diff_filespec_is_binary(two)) )) {
1958 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1959 die("unable to read files to diff");
1960 /* Quite common confusing case */
1961 if (mf1.size == mf2.size &&
1962 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
1963 if (must_show_header)
1964 fprintf(o->file, "%s", header.buf);
1965 goto free_ab_and_return;
1966 }
1967 fprintf(o->file, "%s", header.buf);
1968 strbuf_reset(&header);
1969 if (DIFF_OPT_TST(o, BINARY))
1970 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
1971 else
1972 fprintf(o->file, "%sBinary files %s and %s differ\n",
1973 line_prefix, lbl[0], lbl[1]);
1974 o->found_changes = 1;
1975 }
1976 else {
1977 /* Crazy xdl interfaces.. */
1978 const char *diffopts = getenv("GIT_DIFF_OPTS");
1979 xpparam_t xpp;
1980 xdemitconf_t xecfg;
1981 struct emit_callback ecbdata;
1982 const struct userdiff_funcname *pe;
1983
1984 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
1985 fprintf(o->file, "%s", header.buf);
1986 strbuf_reset(&header);
1987 }
1988
1989 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
1990 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
1991
1992 pe = diff_funcname_pattern(one);
1993 if (!pe)
1994 pe = diff_funcname_pattern(two);
1995
1996 memset(&xpp, 0, sizeof(xpp));
1997 memset(&xecfg, 0, sizeof(xecfg));
1998 memset(&ecbdata, 0, sizeof(ecbdata));
1999 ecbdata.label_path = lbl;
2000 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2001 ecbdata.found_changesp = &o->found_changes;
2002 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2003 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2004 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2005 ecbdata.opt = o;
2006 ecbdata.header = header.len ? &header : NULL;
2007 xpp.flags = o->xdl_opts;
2008 xecfg.ctxlen = o->context;
2009 xecfg.interhunkctxlen = o->interhunkcontext;
2010 xecfg.flags = XDL_EMIT_FUNCNAMES;
2011 if (pe)
2012 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2013 if (!diffopts)
2014 ;
2015 else if (!prefixcmp(diffopts, "--unified="))
2016 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2017 else if (!prefixcmp(diffopts, "-u"))
2018 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2019 if (o->word_diff) {
2020 int i;
2021
2022 ecbdata.diff_words =
2023 xcalloc(1, sizeof(struct diff_words_data));
2024 ecbdata.diff_words->type = o->word_diff;
2025 ecbdata.diff_words->opt = o;
2026 if (!o->word_regex)
2027 o->word_regex = userdiff_word_regex(one);
2028 if (!o->word_regex)
2029 o->word_regex = userdiff_word_regex(two);
2030 if (!o->word_regex)
2031 o->word_regex = diff_word_regex_cfg;
2032 if (o->word_regex) {
2033 ecbdata.diff_words->word_regex = (regex_t *)
2034 xmalloc(sizeof(regex_t));
2035 if (regcomp(ecbdata.diff_words->word_regex,
2036 o->word_regex,
2037 REG_EXTENDED | REG_NEWLINE))
2038 die ("Invalid regular expression: %s",
2039 o->word_regex);
2040 }
2041 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2042 if (o->word_diff == diff_words_styles[i].type) {
2043 ecbdata.diff_words->style =
2044 &diff_words_styles[i];
2045 break;
2046 }
2047 }
2048 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2049 struct diff_words_style *st = ecbdata.diff_words->style;
2050 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2051 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2052 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2053 }
2054 }
2055 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2056 &xpp, &xecfg);
2057 if (o->word_diff)
2058 free_diff_words_data(&ecbdata);
2059 if (textconv_one)
2060 free(mf1.ptr);
2061 if (textconv_two)
2062 free(mf2.ptr);
2063 xdiff_clear_find_func(&xecfg);
2064 }
2065
2066 free_ab_and_return:
2067 strbuf_release(&header);
2068 diff_free_filespec_data(one);
2069 diff_free_filespec_data(two);
2070 free(a_one);
2071 free(b_two);
2072 return;
2073 }
2074
2075 static void builtin_diffstat(const char *name_a, const char *name_b,
2076 struct diff_filespec *one,
2077 struct diff_filespec *two,
2078 struct diffstat_t *diffstat,
2079 struct diff_options *o,
2080 int complete_rewrite)
2081 {
2082 mmfile_t mf1, mf2;
2083 struct diffstat_file *data;
2084
2085 data = diffstat_add(diffstat, name_a, name_b);
2086
2087 if (!one || !two) {
2088 data->is_unmerged = 1;
2089 return;
2090 }
2091
2092 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2093 data->is_binary = 1;
2094 data->added = diff_filespec_size(two);
2095 data->deleted = diff_filespec_size(one);
2096 }
2097
2098 else if (complete_rewrite) {
2099 diff_populate_filespec(one, 0);
2100 diff_populate_filespec(two, 0);
2101 data->deleted = count_lines(one->data, one->size);
2102 data->added = count_lines(two->data, two->size);
2103 }
2104
2105 else {
2106 /* Crazy xdl interfaces.. */
2107 xpparam_t xpp;
2108 xdemitconf_t xecfg;
2109
2110 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2111 die("unable to read files to diff");
2112
2113 memset(&xpp, 0, sizeof(xpp));
2114 memset(&xecfg, 0, sizeof(xecfg));
2115 xpp.flags = o->xdl_opts;
2116 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2117 &xpp, &xecfg);
2118 }
2119
2120 diff_free_filespec_data(one);
2121 diff_free_filespec_data(two);
2122 }
2123
2124 static void builtin_checkdiff(const char *name_a, const char *name_b,
2125 const char *attr_path,
2126 struct diff_filespec *one,
2127 struct diff_filespec *two,
2128 struct diff_options *o)
2129 {
2130 mmfile_t mf1, mf2;
2131 struct checkdiff_t data;
2132
2133 if (!two)
2134 return;
2135
2136 memset(&data, 0, sizeof(data));
2137 data.filename = name_b ? name_b : name_a;
2138 data.lineno = 0;
2139 data.o = o;
2140 data.ws_rule = whitespace_rule(attr_path);
2141 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2142
2143 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2144 die("unable to read files to diff");
2145
2146 /*
2147 * All the other codepaths check both sides, but not checking
2148 * the "old" side here is deliberate. We are checking the newly
2149 * introduced changes, and as long as the "new" side is text, we
2150 * can and should check what it introduces.
2151 */
2152 if (diff_filespec_is_binary(two))
2153 goto free_and_return;
2154 else {
2155 /* Crazy xdl interfaces.. */
2156 xpparam_t xpp;
2157 xdemitconf_t xecfg;
2158
2159 memset(&xpp, 0, sizeof(xpp));
2160 memset(&xecfg, 0, sizeof(xecfg));
2161 xecfg.ctxlen = 1; /* at least one context line */
2162 xpp.flags = 0;
2163 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2164 &xpp, &xecfg);
2165
2166 if (data.ws_rule & WS_BLANK_AT_EOF) {
2167 struct emit_callback ecbdata;
2168 int blank_at_eof;
2169
2170 ecbdata.ws_rule = data.ws_rule;
2171 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2172 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2173
2174 if (blank_at_eof) {
2175 static char *err;
2176 if (!err)
2177 err = whitespace_error_string(WS_BLANK_AT_EOF);
2178 fprintf(o->file, "%s:%d: %s.\n",
2179 data.filename, blank_at_eof, err);
2180 data.status = 1; /* report errors */
2181 }
2182 }
2183 }
2184 free_and_return:
2185 diff_free_filespec_data(one);
2186 diff_free_filespec_data(two);
2187 if (data.status)
2188 DIFF_OPT_SET(o, CHECK_FAILED);
2189 }
2190
2191 struct diff_filespec *alloc_filespec(const char *path)
2192 {
2193 int namelen = strlen(path);
2194 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2195
2196 memset(spec, 0, sizeof(*spec));
2197 spec->path = (char *)(spec + 1);
2198 memcpy(spec->path, path, namelen+1);
2199 spec->count = 1;
2200 spec->is_binary = -1;
2201 return spec;
2202 }
2203
2204 void free_filespec(struct diff_filespec *spec)
2205 {
2206 if (!--spec->count) {
2207 diff_free_filespec_data(spec);
2208 free(spec);
2209 }
2210 }
2211
2212 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2213 unsigned short mode)
2214 {
2215 if (mode) {
2216 spec->mode = canon_mode(mode);
2217 hashcpy(spec->sha1, sha1);
2218 spec->sha1_valid = !is_null_sha1(sha1);
2219 }
2220 }
2221
2222 /*
2223 * Given a name and sha1 pair, if the index tells us the file in
2224 * the work tree has that object contents, return true, so that
2225 * prepare_temp_file() does not have to inflate and extract.
2226 */
2227 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2228 {
2229 struct cache_entry *ce;
2230 struct stat st;
2231 int pos, len;
2232
2233 /*
2234 * We do not read the cache ourselves here, because the
2235 * benchmark with my previous version that always reads cache
2236 * shows that it makes things worse for diff-tree comparing
2237 * two linux-2.6 kernel trees in an already checked out work
2238 * tree. This is because most diff-tree comparisons deal with
2239 * only a small number of files, while reading the cache is
2240 * expensive for a large project, and its cost outweighs the
2241 * savings we get by not inflating the object to a temporary
2242 * file. Practically, this code only helps when we are used
2243 * by diff-cache --cached, which does read the cache before
2244 * calling us.
2245 */
2246 if (!active_cache)
2247 return 0;
2248
2249 /* We want to avoid the working directory if our caller
2250 * doesn't need the data in a normal file, this system
2251 * is rather slow with its stat/open/mmap/close syscalls,
2252 * and the object is contained in a pack file. The pack
2253 * is probably already open and will be faster to obtain
2254 * the data through than the working directory. Loose
2255 * objects however would tend to be slower as they need
2256 * to be individually opened and inflated.
2257 */
2258 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2259 return 0;
2260
2261 len = strlen(name);
2262 pos = cache_name_pos(name, len);
2263 if (pos < 0)
2264 return 0;
2265 ce = active_cache[pos];
2266
2267 /*
2268 * This is not the sha1 we are looking for, or
2269 * unreusable because it is not a regular file.
2270 */
2271 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2272 return 0;
2273
2274 /*
2275 * If ce is marked as "assume unchanged", there is no
2276 * guarantee that work tree matches what we are looking for.
2277 */
2278 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2279 return 0;
2280
2281 /*
2282 * If ce matches the file in the work tree, we can reuse it.
2283 */
2284 if (ce_uptodate(ce) ||
2285 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2286 return 1;
2287
2288 return 0;
2289 }
2290
2291 static int populate_from_stdin(struct diff_filespec *s)
2292 {
2293 struct strbuf buf = STRBUF_INIT;
2294 size_t size = 0;
2295
2296 if (strbuf_read(&buf, 0, 0) < 0)
2297 return error("error while reading from stdin %s",
2298 strerror(errno));
2299
2300 s->should_munmap = 0;
2301 s->data = strbuf_detach(&buf, &size);
2302 s->size = size;
2303 s->should_free = 1;
2304 return 0;
2305 }
2306
2307 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2308 {
2309 int len;
2310 char *data = xmalloc(100), *dirty = "";
2311
2312 /* Are we looking at the work tree? */
2313 if (s->dirty_submodule)
2314 dirty = "-dirty";
2315
2316 len = snprintf(data, 100,
2317 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2318 s->data = data;
2319 s->size = len;
2320 s->should_free = 1;
2321 if (size_only) {
2322 s->data = NULL;
2323 free(data);
2324 }
2325 return 0;
2326 }
2327
2328 /*
2329 * While doing rename detection and pickaxe operation, we may need to
2330 * grab the data for the blob (or file) for our own in-core comparison.
2331 * diff_filespec has data and size fields for this purpose.
2332 */
2333 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2334 {
2335 int err = 0;
2336 if (!DIFF_FILE_VALID(s))
2337 die("internal error: asking to populate invalid file.");
2338 if (S_ISDIR(s->mode))
2339 return -1;
2340
2341 if (s->data)
2342 return 0;
2343
2344 if (size_only && 0 < s->size)
2345 return 0;
2346
2347 if (S_ISGITLINK(s->mode))
2348 return diff_populate_gitlink(s, size_only);
2349
2350 if (!s->sha1_valid ||
2351 reuse_worktree_file(s->path, s->sha1, 0)) {
2352 struct strbuf buf = STRBUF_INIT;
2353 struct stat st;
2354 int fd;
2355
2356 if (!strcmp(s->path, "-"))
2357 return populate_from_stdin(s);
2358
2359 if (lstat(s->path, &st) < 0) {
2360 if (errno == ENOENT) {
2361 err_empty:
2362 err = -1;
2363 empty:
2364 s->data = (char *)"";
2365 s->size = 0;
2366 return err;
2367 }
2368 }
2369 s->size = xsize_t(st.st_size);
2370 if (!s->size)
2371 goto empty;
2372 if (S_ISLNK(st.st_mode)) {
2373 struct strbuf sb = STRBUF_INIT;
2374
2375 if (strbuf_readlink(&sb, s->path, s->size))
2376 goto err_empty;
2377 s->size = sb.len;
2378 s->data = strbuf_detach(&sb, NULL);
2379 s->should_free = 1;
2380 return 0;
2381 }
2382 if (size_only)
2383 return 0;
2384 fd = open(s->path, O_RDONLY);
2385 if (fd < 0)
2386 goto err_empty;
2387 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2388 close(fd);
2389 s->should_munmap = 1;
2390
2391 /*
2392 * Convert from working tree format to canonical git format
2393 */
2394 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2395 size_t size = 0;
2396 munmap(s->data, s->size);
2397 s->should_munmap = 0;
2398 s->data = strbuf_detach(&buf, &size);
2399 s->size = size;
2400 s->should_free = 1;
2401 }
2402 }
2403 else {
2404 enum object_type type;
2405 if (size_only) {
2406 type = sha1_object_info(s->sha1, &s->size);
2407 if (type < 0)
2408 die("unable to read %s", sha1_to_hex(s->sha1));
2409 } else {
2410 s->data = read_sha1_file(s->sha1, &type, &s->size);
2411 if (!s->data)
2412 die("unable to read %s", sha1_to_hex(s->sha1));
2413 s->should_free = 1;
2414 }
2415 }
2416 return 0;
2417 }
2418
2419 void diff_free_filespec_blob(struct diff_filespec *s)
2420 {
2421 if (s->should_free)
2422 free(s->data);
2423 else if (s->should_munmap)
2424 munmap(s->data, s->size);
2425
2426 if (s->should_free || s->should_munmap) {
2427 s->should_free = s->should_munmap = 0;
2428 s->data = NULL;
2429 }
2430 }
2431
2432 void diff_free_filespec_data(struct diff_filespec *s)
2433 {
2434 diff_free_filespec_blob(s);
2435 free(s->cnt_data);
2436 s->cnt_data = NULL;
2437 }
2438
2439 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2440 void *blob,
2441 unsigned long size,
2442 const unsigned char *sha1,
2443 int mode)
2444 {
2445 int fd;
2446 struct strbuf buf = STRBUF_INIT;
2447 struct strbuf template = STRBUF_INIT;
2448 char *path_dup = xstrdup(path);
2449 const char *base = basename(path_dup);
2450
2451 /* Generate "XXXXXX_basename.ext" */
2452 strbuf_addstr(&template, "XXXXXX_");
2453 strbuf_addstr(&template, base);
2454
2455 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2456 strlen(base) + 1);
2457 if (fd < 0)
2458 die_errno("unable to create temp-file");
2459 if (convert_to_working_tree(path,
2460 (const char *)blob, (size_t)size, &buf)) {
2461 blob = buf.buf;
2462 size = buf.len;
2463 }
2464 if (write_in_full(fd, blob, size) != size)
2465 die_errno("unable to write temp-file");
2466 close(fd);
2467 temp->name = temp->tmp_path;
2468 strcpy(temp->hex, sha1_to_hex(sha1));
2469 temp->hex[40] = 0;
2470 sprintf(temp->mode, "%06o", mode);
2471 strbuf_release(&buf);
2472 strbuf_release(&template);
2473 free(path_dup);
2474 }
2475
2476 static struct diff_tempfile *prepare_temp_file(const char *name,
2477 struct diff_filespec *one)
2478 {
2479 struct diff_tempfile *temp = claim_diff_tempfile();
2480
2481 if (!DIFF_FILE_VALID(one)) {
2482 not_a_valid_file:
2483 /* A '-' entry produces this for file-2, and
2484 * a '+' entry produces this for file-1.
2485 */
2486 temp->name = "/dev/null";
2487 strcpy(temp->hex, ".");
2488 strcpy(temp->mode, ".");
2489 return temp;
2490 }
2491
2492 if (!remove_tempfile_installed) {
2493 atexit(remove_tempfile);
2494 sigchain_push_common(remove_tempfile_on_signal);
2495 remove_tempfile_installed = 1;
2496 }
2497
2498 if (!one->sha1_valid ||
2499 reuse_worktree_file(name, one->sha1, 1)) {
2500 struct stat st;
2501 if (lstat(name, &st) < 0) {
2502 if (errno == ENOENT)
2503 goto not_a_valid_file;
2504 die_errno("stat(%s)", name);
2505 }
2506 if (S_ISLNK(st.st_mode)) {
2507 struct strbuf sb = STRBUF_INIT;
2508 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2509 die_errno("readlink(%s)", name);
2510 prep_temp_blob(name, temp, sb.buf, sb.len,
2511 (one->sha1_valid ?
2512 one->sha1 : null_sha1),
2513 (one->sha1_valid ?
2514 one->mode : S_IFLNK));
2515 strbuf_release(&sb);
2516 }
2517 else {
2518 /* we can borrow from the file in the work tree */
2519 temp->name = name;
2520 if (!one->sha1_valid)
2521 strcpy(temp->hex, sha1_to_hex(null_sha1));
2522 else
2523 strcpy(temp->hex, sha1_to_hex(one->sha1));
2524 /* Even though we may sometimes borrow the
2525 * contents from the work tree, we always want
2526 * one->mode. mode is trustworthy even when
2527 * !(one->sha1_valid), as long as
2528 * DIFF_FILE_VALID(one).
2529 */
2530 sprintf(temp->mode, "%06o", one->mode);
2531 }
2532 return temp;
2533 }
2534 else {
2535 if (diff_populate_filespec(one, 0))
2536 die("cannot read data blob for %s", one->path);
2537 prep_temp_blob(name, temp, one->data, one->size,
2538 one->sha1, one->mode);
2539 }
2540 return temp;
2541 }
2542
2543 /* An external diff command takes:
2544 *
2545 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2546 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2547 *
2548 */
2549 static void run_external_diff(const char *pgm,
2550 const char *name,
2551 const char *other,
2552 struct diff_filespec *one,
2553 struct diff_filespec *two,
2554 const char *xfrm_msg,
2555 int complete_rewrite)
2556 {
2557 const char *spawn_arg[10];
2558 int retval;
2559 const char **arg = &spawn_arg[0];
2560
2561 if (one && two) {
2562 struct diff_tempfile *temp_one, *temp_two;
2563 const char *othername = (other ? other : name);
2564 temp_one = prepare_temp_file(name, one);
2565 temp_two = prepare_temp_file(othername, two);
2566 *arg++ = pgm;
2567 *arg++ = name;
2568 *arg++ = temp_one->name;
2569 *arg++ = temp_one->hex;
2570 *arg++ = temp_one->mode;
2571 *arg++ = temp_two->name;
2572 *arg++ = temp_two->hex;
2573 *arg++ = temp_two->mode;
2574 if (other) {
2575 *arg++ = other;
2576 *arg++ = xfrm_msg;
2577 }
2578 } else {
2579 *arg++ = pgm;
2580 *arg++ = name;
2581 }
2582 *arg = NULL;
2583 fflush(NULL);
2584 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2585 remove_tempfile();
2586 if (retval) {
2587 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2588 exit(1);
2589 }
2590 }
2591
2592 static int similarity_index(struct diff_filepair *p)
2593 {
2594 return p->score * 100 / MAX_SCORE;
2595 }
2596
2597 static void fill_metainfo(struct strbuf *msg,
2598 const char *name,
2599 const char *other,
2600 struct diff_filespec *one,
2601 struct diff_filespec *two,
2602 struct diff_options *o,
2603 struct diff_filepair *p,
2604 int *must_show_header,
2605 int use_color)
2606 {
2607 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2608 const char *reset = diff_get_color(use_color, DIFF_RESET);
2609 struct strbuf *msgbuf;
2610 char *line_prefix = "";
2611
2612 *must_show_header = 1;
2613 if (o->output_prefix) {
2614 msgbuf = o->output_prefix(o, o->output_prefix_data);
2615 line_prefix = msgbuf->buf;
2616 }
2617 strbuf_init(msg, PATH_MAX * 2 + 300);
2618 switch (p->status) {
2619 case DIFF_STATUS_COPIED:
2620 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2621 line_prefix, set, similarity_index(p));
2622 strbuf_addf(msg, "%s\n%s%scopy from ",
2623 reset, line_prefix, set);
2624 quote_c_style(name, msg, NULL, 0);
2625 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2626 quote_c_style(other, msg, NULL, 0);
2627 strbuf_addf(msg, "%s\n", reset);
2628 break;
2629 case DIFF_STATUS_RENAMED:
2630 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2631 line_prefix, set, similarity_index(p));
2632 strbuf_addf(msg, "%s\n%s%srename from ",
2633 reset, line_prefix, set);
2634 quote_c_style(name, msg, NULL, 0);
2635 strbuf_addf(msg, "%s\n%s%srename to ",
2636 reset, line_prefix, set);
2637 quote_c_style(other, msg, NULL, 0);
2638 strbuf_addf(msg, "%s\n", reset);
2639 break;
2640 case DIFF_STATUS_MODIFIED:
2641 if (p->score) {
2642 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2643 line_prefix,
2644 set, similarity_index(p), reset);
2645 break;
2646 }
2647 /* fallthru */
2648 default:
2649 *must_show_header = 0;
2650 }
2651 if (one && two && hashcmp(one->sha1, two->sha1)) {
2652 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2653
2654 if (DIFF_OPT_TST(o, BINARY)) {
2655 mmfile_t mf;
2656 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2657 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2658 abbrev = 40;
2659 }
2660 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2661 find_unique_abbrev(one->sha1, abbrev));
2662 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2663 if (one->mode == two->mode)
2664 strbuf_addf(msg, " %06o", one->mode);
2665 strbuf_addf(msg, "%s\n", reset);
2666 }
2667 }
2668
2669 static void run_diff_cmd(const char *pgm,
2670 const char *name,
2671 const char *other,
2672 const char *attr_path,
2673 struct diff_filespec *one,
2674 struct diff_filespec *two,
2675 struct strbuf *msg,
2676 struct diff_options *o,
2677 struct diff_filepair *p)
2678 {
2679 const char *xfrm_msg = NULL;
2680 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2681 int must_show_header = 0;
2682
2683 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2684 pgm = NULL;
2685 else {
2686 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2687 if (drv && drv->external)
2688 pgm = drv->external;
2689 }
2690
2691 if (msg) {
2692 /*
2693 * don't use colors when the header is intended for an
2694 * external diff driver
2695 */
2696 fill_metainfo(msg, name, other, one, two, o, p,
2697 &must_show_header,
2698 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2699 xfrm_msg = msg->len ? msg->buf : NULL;
2700 }
2701
2702 if (pgm) {
2703 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2704 complete_rewrite);
2705 return;
2706 }
2707 if (one && two)
2708 builtin_diff(name, other ? other : name,
2709 one, two, xfrm_msg, must_show_header,
2710 o, complete_rewrite);
2711 else
2712 fprintf(o->file, "* Unmerged path %s\n", name);
2713 }
2714
2715 static void diff_fill_sha1_info(struct diff_filespec *one)
2716 {
2717 if (DIFF_FILE_VALID(one)) {
2718 if (!one->sha1_valid) {
2719 struct stat st;
2720 if (!strcmp(one->path, "-")) {
2721 hashcpy(one->sha1, null_sha1);
2722 return;
2723 }
2724 if (lstat(one->path, &st) < 0)
2725 die_errno("stat '%s'", one->path);
2726 if (index_path(one->sha1, one->path, &st, 0))
2727 die("cannot hash %s", one->path);
2728 }
2729 }
2730 else
2731 hashclr(one->sha1);
2732 }
2733
2734 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2735 {
2736 /* Strip the prefix but do not molest /dev/null and absolute paths */
2737 if (*namep && **namep != '/') {
2738 *namep += prefix_length;
2739 if (**namep == '/')
2740 ++*namep;
2741 }
2742 if (*otherp && **otherp != '/') {
2743 *otherp += prefix_length;
2744 if (**otherp == '/')
2745 ++*otherp;
2746 }
2747 }
2748
2749 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2750 {
2751 const char *pgm = external_diff();
2752 struct strbuf msg;
2753 struct diff_filespec *one = p->one;
2754 struct diff_filespec *two = p->two;
2755 const char *name;
2756 const char *other;
2757 const char *attr_path;
2758
2759 name = p->one->path;
2760 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2761 attr_path = name;
2762 if (o->prefix_length)
2763 strip_prefix(o->prefix_length, &name, &other);
2764
2765 if (DIFF_PAIR_UNMERGED(p)) {
2766 run_diff_cmd(pgm, name, NULL, attr_path,
2767 NULL, NULL, NULL, o, p);
2768 return;
2769 }
2770
2771 diff_fill_sha1_info(one);
2772 diff_fill_sha1_info(two);
2773
2774 if (!pgm &&
2775 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2776 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2777 /*
2778 * a filepair that changes between file and symlink
2779 * needs to be split into deletion and creation.
2780 */
2781 struct diff_filespec *null = alloc_filespec(two->path);
2782 run_diff_cmd(NULL, name, other, attr_path,
2783 one, null, &msg, o, p);
2784 free(null);
2785 strbuf_release(&msg);
2786
2787 null = alloc_filespec(one->path);
2788 run_diff_cmd(NULL, name, other, attr_path,
2789 null, two, &msg, o, p);
2790 free(null);
2791 }
2792 else
2793 run_diff_cmd(pgm, name, other, attr_path,
2794 one, two, &msg, o, p);
2795
2796 strbuf_release(&msg);
2797 }
2798
2799 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2800 struct diffstat_t *diffstat)
2801 {
2802 const char *name;
2803 const char *other;
2804 int complete_rewrite = 0;
2805
2806 if (DIFF_PAIR_UNMERGED(p)) {
2807 /* unmerged */
2808 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2809 return;
2810 }
2811
2812 name = p->one->path;
2813 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2814
2815 if (o->prefix_length)
2816 strip_prefix(o->prefix_length, &name, &other);
2817
2818 diff_fill_sha1_info(p->one);
2819 diff_fill_sha1_info(p->two);
2820
2821 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2822 complete_rewrite = 1;
2823 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2824 }
2825
2826 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2827 {
2828 const char *name;
2829 const char *other;
2830 const char *attr_path;
2831
2832 if (DIFF_PAIR_UNMERGED(p)) {
2833 /* unmerged */
2834 return;
2835 }
2836
2837 name = p->one->path;
2838 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2839 attr_path = other ? other : name;
2840
2841 if (o->prefix_length)
2842 strip_prefix(o->prefix_length, &name, &other);
2843
2844 diff_fill_sha1_info(p->one);
2845 diff_fill_sha1_info(p->two);
2846
2847 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2848 }
2849
2850 void diff_setup(struct diff_options *options)
2851 {
2852 memcpy(options, &default_diff_options, sizeof(*options));
2853
2854 options->file = stdout;
2855
2856 options->line_termination = '\n';
2857 options->break_opt = -1;
2858 options->rename_limit = -1;
2859 options->dirstat_percent = 3;
2860 options->context = 3;
2861
2862 options->change = diff_change;
2863 options->add_remove = diff_addremove;
2864 if (diff_use_color_default > 0)
2865 DIFF_OPT_SET(options, COLOR_DIFF);
2866 options->detect_rename = diff_detect_rename_default;
2867
2868 if (diff_no_prefix) {
2869 options->a_prefix = options->b_prefix = "";
2870 } else if (!diff_mnemonic_prefix) {
2871 options->a_prefix = "a/";
2872 options->b_prefix = "b/";
2873 }
2874 }
2875
2876 int diff_setup_done(struct diff_options *options)
2877 {
2878 int count = 0;
2879
2880 if (options->output_format & DIFF_FORMAT_NAME)
2881 count++;
2882 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2883 count++;
2884 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2885 count++;
2886 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2887 count++;
2888 if (count > 1)
2889 die("--name-only, --name-status, --check and -s are mutually exclusive");
2890
2891 /*
2892 * Most of the time we can say "there are changes"
2893 * only by checking if there are changed paths, but
2894 * --ignore-whitespace* options force us to look
2895 * inside contents.
2896 */
2897
2898 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2899 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2900 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2901 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2902 else
2903 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2904
2905 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2906 options->detect_rename = DIFF_DETECT_COPY;
2907
2908 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2909 options->prefix = NULL;
2910 if (options->prefix)
2911 options->prefix_length = strlen(options->prefix);
2912 else
2913 options->prefix_length = 0;
2914
2915 if (options->output_format & (DIFF_FORMAT_NAME |
2916 DIFF_FORMAT_NAME_STATUS |
2917 DIFF_FORMAT_CHECKDIFF |
2918 DIFF_FORMAT_NO_OUTPUT))
2919 options->output_format &= ~(DIFF_FORMAT_RAW |
2920 DIFF_FORMAT_NUMSTAT |
2921 DIFF_FORMAT_DIFFSTAT |
2922 DIFF_FORMAT_SHORTSTAT |
2923 DIFF_FORMAT_DIRSTAT |
2924 DIFF_FORMAT_SUMMARY |
2925 DIFF_FORMAT_PATCH);
2926
2927 /*
2928 * These cases always need recursive; we do not drop caller-supplied
2929 * recursive bits for other formats here.
2930 */
2931 if (options->output_format & (DIFF_FORMAT_PATCH |
2932 DIFF_FORMAT_NUMSTAT |
2933 DIFF_FORMAT_DIFFSTAT |
2934 DIFF_FORMAT_SHORTSTAT |
2935 DIFF_FORMAT_DIRSTAT |
2936 DIFF_FORMAT_SUMMARY |
2937 DIFF_FORMAT_CHECKDIFF))
2938 DIFF_OPT_SET(options, RECURSIVE);
2939 /*
2940 * Also pickaxe would not work very well if you do not say recursive
2941 */
2942 if (options->pickaxe)
2943 DIFF_OPT_SET(options, RECURSIVE);
2944 /*
2945 * When patches are generated, submodules diffed against the work tree
2946 * must be checked for dirtiness too so it can be shown in the output
2947 */
2948 if (options->output_format & DIFF_FORMAT_PATCH)
2949 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2950
2951 if (options->detect_rename && options->rename_limit < 0)
2952 options->rename_limit = diff_rename_limit_default;
2953 if (options->setup & DIFF_SETUP_USE_CACHE) {
2954 if (!active_cache)
2955 /* read-cache does not die even when it fails
2956 * so it is safe for us to do this here. Also
2957 * it does not smudge active_cache or active_nr
2958 * when it fails, so we do not have to worry about
2959 * cleaning it up ourselves either.
2960 */
2961 read_cache();
2962 }
2963 if (options->abbrev <= 0 || 40 < options->abbrev)
2964 options->abbrev = 40; /* full */
2965
2966 /*
2967 * It does not make sense to show the first hit we happened
2968 * to have found. It does not make sense not to return with
2969 * exit code in such a case either.
2970 */
2971 if (DIFF_OPT_TST(options, QUICK)) {
2972 options->output_format = DIFF_FORMAT_NO_OUTPUT;
2973 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
2974 }
2975
2976 return 0;
2977 }
2978
2979 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
2980 {
2981 char c, *eq;
2982 int len;
2983
2984 if (*arg != '-')
2985 return 0;
2986 c = *++arg;
2987 if (!c)
2988 return 0;
2989 if (c == arg_short) {
2990 c = *++arg;
2991 if (!c)
2992 return 1;
2993 if (val && isdigit(c)) {
2994 char *end;
2995 int n = strtoul(arg, &end, 10);
2996 if (*end)
2997 return 0;
2998 *val = n;
2999 return 1;
3000 }
3001 return 0;
3002 }
3003 if (c != '-')
3004 return 0;
3005 arg++;
3006 eq = strchr(arg, '=');
3007 if (eq)
3008 len = eq - arg;
3009 else
3010 len = strlen(arg);
3011 if (!len || strncmp(arg, arg_long, len))
3012 return 0;
3013 if (eq) {
3014 int n;
3015 char *end;
3016 if (!isdigit(*++eq))
3017 return 0;
3018 n = strtoul(eq, &end, 10);
3019 if (*end)
3020 return 0;
3021 *val = n;
3022 }
3023 return 1;
3024 }
3025
3026 static int diff_scoreopt_parse(const char *opt);
3027
3028 static inline int short_opt(char opt, const char **argv,
3029 const char **optarg)
3030 {
3031 const char *arg = argv[0];
3032 if (arg[0] != '-' || arg[1] != opt)
3033 return 0;
3034 if (arg[2] != '\0') {
3035 *optarg = arg + 2;
3036 return 1;
3037 }
3038 if (!argv[1])
3039 die("Option '%c' requires a value", opt);
3040 *optarg = argv[1];
3041 return 2;
3042 }
3043
3044 int parse_long_opt(const char *opt, const char **argv,
3045 const char **optarg)
3046 {
3047 const char *arg = argv[0];
3048 if (arg[0] != '-' || arg[1] != '-')
3049 return 0;
3050 arg += strlen("--");
3051 if (prefixcmp(arg, opt))
3052 return 0;
3053 arg += strlen(opt);
3054 if (*arg == '=') { /* sticked form: --option=value */
3055 *optarg = arg + 1;
3056 return 1;
3057 }
3058 if (*arg != '\0')
3059 return 0;
3060 /* separate form: --option value */
3061 if (!argv[1])
3062 die("Option '--%s' requires a value", opt);
3063 *optarg = argv[1];
3064 return 2;
3065 }
3066
3067 static int stat_opt(struct diff_options *options, const char **av)
3068 {
3069 const char *arg = av[0];
3070 char *end;
3071 int width = options->stat_width;
3072 int name_width = options->stat_name_width;
3073 int argcount = 1;
3074
3075 arg += strlen("--stat");
3076 end = (char *)arg;
3077
3078 switch (*arg) {
3079 case '-':
3080 if (!prefixcmp(arg, "-width")) {
3081 arg += strlen("-width");
3082 if (*arg == '=')
3083 width = strtoul(arg + 1, &end, 10);
3084 else if (!*arg && !av[1])
3085 die("Option '--stat-width' requires a value");
3086 else if (!*arg) {
3087 width = strtoul(av[1], &end, 10);
3088 argcount = 2;
3089 }
3090 } else if (!prefixcmp(arg, "-name-width")) {
3091 arg += strlen("-name-width");
3092 if (*arg == '=')
3093 name_width = strtoul(arg + 1, &end, 10);
3094 else if (!*arg && !av[1])
3095 die("Option '--stat-name-width' requires a value");
3096 else if (!*arg) {
3097 name_width = strtoul(av[1], &end, 10);
3098 argcount = 2;
3099 }
3100 }
3101 break;
3102 case '=':
3103 width = strtoul(arg+1, &end, 10);
3104 if (*end == ',')
3105 name_width = strtoul(end+1, &end, 10);
3106 }
3107
3108 /* Important! This checks all the error cases! */
3109 if (*end)
3110 return 0;
3111 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3112 options->stat_name_width = name_width;
3113 options->stat_width = width;
3114 return argcount;
3115 }
3116
3117 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3118 {
3119 const char *arg = av[0];
3120 const char *optarg;
3121 int argcount;
3122
3123 /* Output format options */
3124 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3125 options->output_format |= DIFF_FORMAT_PATCH;
3126 else if (opt_arg(arg, 'U', "unified", &options->context))
3127 options->output_format |= DIFF_FORMAT_PATCH;
3128 else if (!strcmp(arg, "--raw"))
3129 options->output_format |= DIFF_FORMAT_RAW;
3130 else if (!strcmp(arg, "--patch-with-raw"))
3131 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3132 else if (!strcmp(arg, "--numstat"))
3133 options->output_format |= DIFF_FORMAT_NUMSTAT;
3134 else if (!strcmp(arg, "--shortstat"))
3135 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3136 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3137 options->output_format |= DIFF_FORMAT_DIRSTAT;
3138 else if (!strcmp(arg, "--cumulative")) {
3139 options->output_format |= DIFF_FORMAT_DIRSTAT;
3140 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3141 } else if (opt_arg(arg, 0, "dirstat-by-file",
3142 &options->dirstat_percent)) {
3143 options->output_format |= DIFF_FORMAT_DIRSTAT;
3144 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3145 }
3146 else if (!strcmp(arg, "--check"))
3147 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3148 else if (!strcmp(arg, "--summary"))
3149 options->output_format |= DIFF_FORMAT_SUMMARY;
3150 else if (!strcmp(arg, "--patch-with-stat"))
3151 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3152 else if (!strcmp(arg, "--name-only"))
3153 options->output_format |= DIFF_FORMAT_NAME;
3154 else if (!strcmp(arg, "--name-status"))
3155 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3156 else if (!strcmp(arg, "-s"))
3157 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3158 else if (!prefixcmp(arg, "--stat"))
3159 /* --stat, --stat-width, or --stat-name-width */
3160 return stat_opt(options, av);
3161
3162 /* renames options */
3163 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3164 !strcmp(arg, "--break-rewrites")) {
3165 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3166 return error("invalid argument to -B: %s", arg+2);
3167 }
3168 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3169 !strcmp(arg, "--find-renames")) {
3170 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3171 return error("invalid argument to -M: %s", arg+2);
3172 options->detect_rename = DIFF_DETECT_RENAME;
3173 }
3174 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3175 !strcmp(arg, "--find-copies")) {
3176 if (options->detect_rename == DIFF_DETECT_COPY)
3177 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3178 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3179 return error("invalid argument to -C: %s", arg+2);
3180 options->detect_rename = DIFF_DETECT_COPY;
3181 }
3182 else if (!strcmp(arg, "--no-renames"))
3183 options->detect_rename = 0;
3184 else if (!strcmp(arg, "--relative"))
3185 DIFF_OPT_SET(options, RELATIVE_NAME);
3186 else if (!prefixcmp(arg, "--relative=")) {
3187 DIFF_OPT_SET(options, RELATIVE_NAME);
3188 options->prefix = arg + 11;
3189 }
3190
3191 /* xdiff options */
3192 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3193 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3194 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3195 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3196 else if (!strcmp(arg, "--ignore-space-at-eol"))
3197 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3198 else if (!strcmp(arg, "--patience"))
3199 DIFF_XDL_SET(options, PATIENCE_DIFF);
3200
3201 /* flags options */
3202 else if (!strcmp(arg, "--binary")) {
3203 options->output_format |= DIFF_FORMAT_PATCH;
3204 DIFF_OPT_SET(options, BINARY);
3205 }
3206 else if (!strcmp(arg, "--full-index"))
3207 DIFF_OPT_SET(options, FULL_INDEX);
3208 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3209 DIFF_OPT_SET(options, TEXT);
3210 else if (!strcmp(arg, "-R"))
3211 DIFF_OPT_SET(options, REVERSE_DIFF);
3212 else if (!strcmp(arg, "--find-copies-harder"))
3213 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3214 else if (!strcmp(arg, "--follow"))
3215 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3216 else if (!strcmp(arg, "--color"))
3217 DIFF_OPT_SET(options, COLOR_DIFF);
3218 else if (!prefixcmp(arg, "--color=")) {
3219 int value = git_config_colorbool(NULL, arg+8, -1);
3220 if (value == 0)
3221 DIFF_OPT_CLR(options, COLOR_DIFF);
3222 else if (value > 0)
3223 DIFF_OPT_SET(options, COLOR_DIFF);
3224 else
3225 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3226 }
3227 else if (!strcmp(arg, "--no-color"))
3228 DIFF_OPT_CLR(options, COLOR_DIFF);
3229 else if (!strcmp(arg, "--color-words")) {
3230 DIFF_OPT_SET(options, COLOR_DIFF);
3231 options->word_diff = DIFF_WORDS_COLOR;
3232 }
3233 else if (!prefixcmp(arg, "--color-words=")) {
3234 DIFF_OPT_SET(options, COLOR_DIFF);
3235 options->word_diff = DIFF_WORDS_COLOR;
3236 options->word_regex = arg + 14;
3237 }
3238 else if (!strcmp(arg, "--word-diff")) {
3239 if (options->word_diff == DIFF_WORDS_NONE)
3240 options->word_diff = DIFF_WORDS_PLAIN;
3241 }
3242 else if (!prefixcmp(arg, "--word-diff=")) {
3243 const char *type = arg + 12;
3244 if (!strcmp(type, "plain"))
3245 options->word_diff = DIFF_WORDS_PLAIN;
3246 else if (!strcmp(type, "color")) {
3247 DIFF_OPT_SET(options, COLOR_DIFF);
3248 options->word_diff = DIFF_WORDS_COLOR;
3249 }
3250 else if (!strcmp(type, "porcelain"))
3251 options->word_diff = DIFF_WORDS_PORCELAIN;
3252 else if (!strcmp(type, "none"))
3253 options->word_diff = DIFF_WORDS_NONE;
3254 else
3255 die("bad --word-diff argument: %s", type);
3256 }
3257 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3258 if (options->word_diff == DIFF_WORDS_NONE)
3259 options->word_diff = DIFF_WORDS_PLAIN;
3260 options->word_regex = optarg;
3261 return argcount;
3262 }
3263 else if (!strcmp(arg, "--exit-code"))
3264 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3265 else if (!strcmp(arg, "--quiet"))
3266 DIFF_OPT_SET(options, QUICK);
3267 else if (!strcmp(arg, "--ext-diff"))
3268 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3269 else if (!strcmp(arg, "--no-ext-diff"))
3270 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3271 else if (!strcmp(arg, "--textconv"))
3272 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3273 else if (!strcmp(arg, "--no-textconv"))
3274 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3275 else if (!strcmp(arg, "--ignore-submodules")) {
3276 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3277 handle_ignore_submodules_arg(options, "all");
3278 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3279 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3280 handle_ignore_submodules_arg(options, arg + 20);
3281 } else if (!strcmp(arg, "--submodule"))
3282 DIFF_OPT_SET(options, SUBMODULE_LOG);
3283 else if (!prefixcmp(arg, "--submodule=")) {
3284 if (!strcmp(arg + 12, "log"))
3285 DIFF_OPT_SET(options, SUBMODULE_LOG);
3286 }
3287
3288 /* misc options */
3289 else if (!strcmp(arg, "-z"))
3290 options->line_termination = 0;
3291 else if ((argcount = short_opt('l', av, &optarg))) {
3292 options->rename_limit = strtoul(optarg, NULL, 10);
3293 return argcount;
3294 }
3295 else if ((argcount = short_opt('S', av, &optarg))) {
3296 options->pickaxe = optarg;
3297 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3298 return argcount;
3299 } else if ((argcount = short_opt('G', av, &optarg))) {
3300 options->pickaxe = optarg;
3301 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3302 return argcount;
3303 }
3304 else if (!strcmp(arg, "--pickaxe-all"))
3305 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3306 else if (!strcmp(arg, "--pickaxe-regex"))
3307 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3308 else if ((argcount = short_opt('O', av, &optarg))) {
3309 options->orderfile = optarg;
3310 return argcount;
3311 }
3312 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3313 options->filter = optarg;
3314 return argcount;
3315 }
3316 else if (!strcmp(arg, "--abbrev"))
3317 options->abbrev = DEFAULT_ABBREV;
3318 else if (!prefixcmp(arg, "--abbrev=")) {
3319 options->abbrev = strtoul(arg + 9, NULL, 10);
3320 if (options->abbrev < MINIMUM_ABBREV)
3321 options->abbrev = MINIMUM_ABBREV;
3322 else if (40 < options->abbrev)
3323 options->abbrev = 40;
3324 }
3325 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3326 options->a_prefix = optarg;
3327 return argcount;
3328 }
3329 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3330 options->b_prefix = optarg;
3331 return argcount;
3332 }
3333 else if (!strcmp(arg, "--no-prefix"))
3334 options->a_prefix = options->b_prefix = "";
3335 else if (opt_arg(arg, '\0', "inter-hunk-context",
3336 &options->interhunkcontext))
3337 ;
3338 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3339 options->file = fopen(optarg, "w");
3340 if (!options->file)
3341 die_errno("Could not open '%s'", optarg);
3342 options->close_file = 1;
3343 return argcount;
3344 } else
3345 return 0;
3346 return 1;
3347 }
3348
3349 int parse_rename_score(const char **cp_p)
3350 {
3351 unsigned long num, scale;
3352 int ch, dot;
3353 const char *cp = *cp_p;
3354
3355 num = 0;
3356 scale = 1;
3357 dot = 0;
3358 for (;;) {
3359 ch = *cp;
3360 if ( !dot && ch == '.' ) {
3361 scale = 1;
3362 dot = 1;
3363 } else if ( ch == '%' ) {
3364 scale = dot ? scale*100 : 100;
3365 cp++; /* % is always at the end */
3366 break;
3367 } else if ( ch >= '0' && ch <= '9' ) {
3368 if ( scale < 100000 ) {
3369 scale *= 10;
3370 num = (num*10) + (ch-'0');
3371 }
3372 } else {
3373 break;
3374 }
3375 cp++;
3376 }
3377 *cp_p = cp;
3378
3379 /* user says num divided by scale and we say internally that
3380 * is MAX_SCORE * num / scale.
3381 */
3382 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3383 }
3384
3385 static int diff_scoreopt_parse(const char *opt)
3386 {
3387 int opt1, opt2, cmd;
3388
3389 if (*opt++ != '-')
3390 return -1;
3391 cmd = *opt++;
3392 if (cmd == '-') {
3393 /* convert the long-form arguments into short-form versions */
3394 if (!prefixcmp(opt, "break-rewrites")) {
3395 opt += strlen("break-rewrites");
3396 if (*opt == 0 || *opt++ == '=')
3397 cmd = 'B';
3398 } else if (!prefixcmp(opt, "find-copies")) {
3399 opt += strlen("find-copies");
3400 if (*opt == 0 || *opt++ == '=')
3401 cmd = 'C';
3402 } else if (!prefixcmp(opt, "find-renames")) {
3403 opt += strlen("find-renames");
3404 if (*opt == 0 || *opt++ == '=')
3405 cmd = 'M';
3406 }
3407 }
3408 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3409 return -1; /* that is not a -M, -C nor -B option */
3410
3411 opt1 = parse_rename_score(&opt);
3412 if (cmd != 'B')
3413 opt2 = 0;
3414 else {
3415 if (*opt == 0)
3416 opt2 = 0;
3417 else if (*opt != '/')
3418 return -1; /* we expect -B80/99 or -B80 */
3419 else {
3420 opt++;
3421 opt2 = parse_rename_score(&opt);
3422 }
3423 }
3424 if (*opt != 0)
3425 return -1;
3426 return opt1 | (opt2 << 16);
3427 }
3428
3429 struct diff_queue_struct diff_queued_diff;
3430
3431 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3432 {
3433 if (queue->alloc <= queue->nr) {
3434 queue->alloc = alloc_nr(queue->alloc);
3435 queue->queue = xrealloc(queue->queue,
3436 sizeof(dp) * queue->alloc);
3437 }
3438 queue->queue[queue->nr++] = dp;
3439 }
3440
3441 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3442 struct diff_filespec *one,
3443 struct diff_filespec *two)
3444 {
3445 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3446 dp->one = one;
3447 dp->two = two;
3448 if (queue)
3449 diff_q(queue, dp);
3450 return dp;
3451 }
3452
3453 void diff_free_filepair(struct diff_filepair *p)
3454 {
3455 free_filespec(p->one);
3456 free_filespec(p->two);
3457 free(p);
3458 }
3459
3460 /* This is different from find_unique_abbrev() in that
3461 * it stuffs the result with dots for alignment.
3462 */
3463 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3464 {
3465 int abblen;
3466 const char *abbrev;
3467 if (len == 40)
3468 return sha1_to_hex(sha1);
3469
3470 abbrev = find_unique_abbrev(sha1, len);
3471 abblen = strlen(abbrev);
3472 if (abblen < 37) {
3473 static char hex[41];
3474 if (len < abblen && abblen <= len + 2)
3475 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3476 else
3477 sprintf(hex, "%s...", abbrev);
3478 return hex;
3479 }
3480 return sha1_to_hex(sha1);
3481 }
3482
3483 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3484 {
3485 int line_termination = opt->line_termination;
3486 int inter_name_termination = line_termination ? '\t' : '\0';
3487 if (opt->output_prefix) {
3488 struct strbuf *msg = NULL;
3489 msg = opt->output_prefix(opt, opt->output_prefix_data);
3490 fprintf(opt->file, "%s", msg->buf);
3491 }
3492
3493 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3494 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3495 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3496 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3497 }
3498 if (p->score) {
3499 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3500 inter_name_termination);
3501 } else {
3502 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3503 }
3504
3505 if (p->status == DIFF_STATUS_COPIED ||
3506 p->status == DIFF_STATUS_RENAMED) {
3507 const char *name_a, *name_b;
3508 name_a = p->one->path;
3509 name_b = p->two->path;
3510 strip_prefix(opt->prefix_length, &name_a, &name_b);
3511 write_name_quoted(name_a, opt->file, inter_name_termination);
3512 write_name_quoted(name_b, opt->file, line_termination);
3513 } else {
3514 const char *name_a, *name_b;
3515 name_a = p->one->mode ? p->one->path : p->two->path;
3516 name_b = NULL;
3517 strip_prefix(opt->prefix_length, &name_a, &name_b);
3518 write_name_quoted(name_a, opt->file, line_termination);
3519 }
3520 }
3521
3522 int diff_unmodified_pair(struct diff_filepair *p)
3523 {
3524 /* This function is written stricter than necessary to support
3525 * the currently implemented transformers, but the idea is to
3526 * let transformers to produce diff_filepairs any way they want,
3527 * and filter and clean them up here before producing the output.
3528 */
3529 struct diff_filespec *one = p->one, *two = p->two;
3530
3531 if (DIFF_PAIR_UNMERGED(p))
3532 return 0; /* unmerged is interesting */
3533
3534 /* deletion, addition, mode or type change
3535 * and rename are all interesting.
3536 */
3537 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3538 DIFF_PAIR_MODE_CHANGED(p) ||
3539 strcmp(one->path, two->path))
3540 return 0;
3541
3542 /* both are valid and point at the same path. that is, we are
3543 * dealing with a change.
3544 */
3545 if (one->sha1_valid && two->sha1_valid &&
3546 !hashcmp(one->sha1, two->sha1) &&
3547 !one->dirty_submodule && !two->dirty_submodule)
3548 return 1; /* no change */
3549 if (!one->sha1_valid && !two->sha1_valid)
3550 return 1; /* both look at the same file on the filesystem. */
3551 return 0;
3552 }
3553
3554 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3555 {
3556 if (diff_unmodified_pair(p))
3557 return;
3558
3559 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3560 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3561 return; /* no tree diffs in patch format */
3562
3563 run_diff(p, o);
3564 }
3565
3566 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3567 struct diffstat_t *diffstat)
3568 {
3569 if (diff_unmodified_pair(p))
3570 return;
3571
3572 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3573 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3574 return; /* no useful stat for tree diffs */
3575
3576 run_diffstat(p, o, diffstat);
3577 }
3578
3579 static void diff_flush_checkdiff(struct diff_filepair *p,
3580 struct diff_options *o)
3581 {
3582 if (diff_unmodified_pair(p))
3583 return;
3584
3585 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3586 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3587 return; /* nothing to check in tree diffs */
3588
3589 run_checkdiff(p, o);
3590 }
3591
3592 int diff_queue_is_empty(void)
3593 {
3594 struct diff_queue_struct *q = &diff_queued_diff;
3595 int i;
3596 for (i = 0; i < q->nr; i++)
3597 if (!diff_unmodified_pair(q->queue[i]))
3598 return 0;
3599 return 1;
3600 }
3601
3602 #if DIFF_DEBUG
3603 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3604 {
3605 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3606 x, one ? one : "",
3607 s->path,
3608 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3609 s->mode,
3610 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3611 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3612 x, one ? one : "",
3613 s->size, s->xfrm_flags);
3614 }
3615
3616 void diff_debug_filepair(const struct diff_filepair *p, int i)
3617 {
3618 diff_debug_filespec(p->one, i, "one");
3619 diff_debug_filespec(p->two, i, "two");
3620 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3621 p->score, p->status ? p->status : '?',
3622 p->one->rename_used, p->broken_pair);
3623 }
3624
3625 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3626 {
3627 int i;
3628 if (msg)
3629 fprintf(stderr, "%s\n", msg);
3630 fprintf(stderr, "q->nr = %d\n", q->nr);
3631 for (i = 0; i < q->nr; i++) {
3632 struct diff_filepair *p = q->queue[i];
3633 diff_debug_filepair(p, i);
3634 }
3635 }
3636 #endif
3637
3638 static void diff_resolve_rename_copy(void)
3639 {
3640 int i;
3641 struct diff_filepair *p;
3642 struct diff_queue_struct *q = &diff_queued_diff;
3643
3644 diff_debug_queue("resolve-rename-copy", q);
3645
3646 for (i = 0; i < q->nr; i++) {
3647 p = q->queue[i];
3648 p->status = 0; /* undecided */
3649 if (DIFF_PAIR_UNMERGED(p))
3650 p->status = DIFF_STATUS_UNMERGED;
3651 else if (!DIFF_FILE_VALID(p->one))
3652 p->status = DIFF_STATUS_ADDED;
3653 else if (!DIFF_FILE_VALID(p->two))
3654 p->status = DIFF_STATUS_DELETED;
3655 else if (DIFF_PAIR_TYPE_CHANGED(p))
3656 p->status = DIFF_STATUS_TYPE_CHANGED;
3657
3658 /* from this point on, we are dealing with a pair
3659 * whose both sides are valid and of the same type, i.e.
3660 * either in-place edit or rename/copy edit.
3661 */
3662 else if (DIFF_PAIR_RENAME(p)) {
3663 /*
3664 * A rename might have re-connected a broken
3665 * pair up, causing the pathnames to be the
3666 * same again. If so, that's not a rename at
3667 * all, just a modification..
3668 *
3669 * Otherwise, see if this source was used for
3670 * multiple renames, in which case we decrement
3671 * the count, and call it a copy.
3672 */
3673 if (!strcmp(p->one->path, p->two->path))
3674 p->status = DIFF_STATUS_MODIFIED;
3675 else if (--p->one->rename_used > 0)
3676 p->status = DIFF_STATUS_COPIED;
3677 else
3678 p->status = DIFF_STATUS_RENAMED;
3679 }
3680 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3681 p->one->mode != p->two->mode ||
3682 p->one->dirty_submodule ||
3683 p->two->dirty_submodule ||
3684 is_null_sha1(p->one->sha1))
3685 p->status = DIFF_STATUS_MODIFIED;
3686 else {
3687 /* This is a "no-change" entry and should not
3688 * happen anymore, but prepare for broken callers.
3689 */
3690 error("feeding unmodified %s to diffcore",
3691 p->one->path);
3692 p->status = DIFF_STATUS_UNKNOWN;
3693 }
3694 }
3695 diff_debug_queue("resolve-rename-copy done", q);
3696 }
3697
3698 static int check_pair_status(struct diff_filepair *p)
3699 {
3700 switch (p->status) {
3701 case DIFF_STATUS_UNKNOWN:
3702 return 0;
3703 case 0:
3704 die("internal error in diff-resolve-rename-copy");
3705 default:
3706 return 1;
3707 }
3708 }
3709
3710 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3711 {
3712 int fmt = opt->output_format;
3713
3714 if (fmt & DIFF_FORMAT_CHECKDIFF)
3715 diff_flush_checkdiff(p, opt);
3716 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3717 diff_flush_raw(p, opt);
3718 else if (fmt & DIFF_FORMAT_NAME) {
3719 const char *name_a, *name_b;
3720 name_a = p->two->path;
3721 name_b = NULL;
3722 strip_prefix(opt->prefix_length, &name_a, &name_b);
3723 write_name_quoted(name_a, opt->file, opt->line_termination);
3724 }
3725 }
3726
3727 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3728 {
3729 if (fs->mode)
3730 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3731 else
3732 fprintf(file, " %s ", newdelete);
3733 write_name_quoted(fs->path, file, '\n');
3734 }
3735
3736
3737 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3738 const char *line_prefix)
3739 {
3740 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3741 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3742 p->two->mode, show_name ? ' ' : '\n');
3743 if (show_name) {
3744 write_name_quoted(p->two->path, file, '\n');
3745 }
3746 }
3747 }
3748
3749 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3750 const char *line_prefix)
3751 {
3752 char *names = pprint_rename(p->one->path, p->two->path);
3753
3754 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3755 free(names);
3756 show_mode_change(file, p, 0, line_prefix);
3757 }
3758
3759 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3760 {
3761 FILE *file = opt->file;
3762 char *line_prefix = "";
3763
3764 if (opt->output_prefix) {
3765 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3766 line_prefix = buf->buf;
3767 }
3768
3769 switch(p->status) {
3770 case DIFF_STATUS_DELETED:
3771 fputs(line_prefix, file);
3772 show_file_mode_name(file, "delete", p->one);
3773 break;
3774 case DIFF_STATUS_ADDED:
3775 fputs(line_prefix, file);
3776 show_file_mode_name(file, "create", p->two);
3777 break;
3778 case DIFF_STATUS_COPIED:
3779 fputs(line_prefix, file);
3780 show_rename_copy(file, "copy", p, line_prefix);
3781 break;
3782 case DIFF_STATUS_RENAMED:
3783 fputs(line_prefix, file);
3784 show_rename_copy(file, "rename", p, line_prefix);
3785 break;
3786 default:
3787 if (p->score) {
3788 fprintf(file, "%s rewrite ", line_prefix);
3789 write_name_quoted(p->two->path, file, ' ');
3790 fprintf(file, "(%d%%)\n", similarity_index(p));
3791 }
3792 show_mode_change(file, p, !p->score, line_prefix);
3793 break;
3794 }
3795 }
3796
3797 struct patch_id_t {
3798 git_SHA_CTX *ctx;
3799 int patchlen;
3800 };
3801
3802 static int remove_space(char *line, int len)
3803 {
3804 int i;
3805 char *dst = line;
3806 unsigned char c;
3807
3808 for (i = 0; i < len; i++)
3809 if (!isspace((c = line[i])))
3810 *dst++ = c;
3811
3812 return dst - line;
3813 }
3814
3815 static void patch_id_consume(void *priv, char *line, unsigned long len)
3816 {
3817 struct patch_id_t *data = priv;
3818 int new_len;
3819
3820 /* Ignore line numbers when computing the SHA1 of the patch */
3821 if (!prefixcmp(line, "@@ -"))
3822 return;
3823
3824 new_len = remove_space(line, len);
3825
3826 git_SHA1_Update(data->ctx, line, new_len);
3827 data->patchlen += new_len;
3828 }
3829
3830 /* returns 0 upon success, and writes result into sha1 */
3831 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3832 {
3833 struct diff_queue_struct *q = &diff_queued_diff;
3834 int i;
3835 git_SHA_CTX ctx;
3836 struct patch_id_t data;
3837 char buffer[PATH_MAX * 4 + 20];
3838
3839 git_SHA1_Init(&ctx);
3840 memset(&data, 0, sizeof(struct patch_id_t));
3841 data.ctx = &ctx;
3842
3843 for (i = 0; i < q->nr; i++) {
3844 xpparam_t xpp;
3845 xdemitconf_t xecfg;
3846 mmfile_t mf1, mf2;
3847 struct diff_filepair *p = q->queue[i];
3848 int len1, len2;
3849
3850 memset(&xpp, 0, sizeof(xpp));
3851 memset(&xecfg, 0, sizeof(xecfg));
3852 if (p->status == 0)
3853 return error("internal diff status error");
3854 if (p->status == DIFF_STATUS_UNKNOWN)
3855 continue;
3856 if (diff_unmodified_pair(p))
3857 continue;
3858 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3859 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3860 continue;
3861 if (DIFF_PAIR_UNMERGED(p))
3862 continue;
3863
3864 diff_fill_sha1_info(p->one);
3865 diff_fill_sha1_info(p->two);
3866 if (fill_mmfile(&mf1, p->one) < 0 ||
3867 fill_mmfile(&mf2, p->two) < 0)
3868 return error("unable to read files to diff");
3869
3870 len1 = remove_space(p->one->path, strlen(p->one->path));
3871 len2 = remove_space(p->two->path, strlen(p->two->path));
3872 if (p->one->mode == 0)
3873 len1 = snprintf(buffer, sizeof(buffer),
3874 "diff--gita/%.*sb/%.*s"
3875 "newfilemode%06o"
3876 "---/dev/null"
3877 "+++b/%.*s",
3878 len1, p->one->path,
3879 len2, p->two->path,
3880 p->two->mode,
3881 len2, p->two->path);
3882 else if (p->two->mode == 0)
3883 len1 = snprintf(buffer, sizeof(buffer),
3884 "diff--gita/%.*sb/%.*s"
3885 "deletedfilemode%06o"
3886 "---a/%.*s"
3887 "+++/dev/null",
3888 len1, p->one->path,
3889 len2, p->two->path,
3890 p->one->mode,
3891 len1, p->one->path);
3892 else
3893 len1 = snprintf(buffer, sizeof(buffer),
3894 "diff--gita/%.*sb/%.*s"
3895 "---a/%.*s"
3896 "+++b/%.*s",
3897 len1, p->one->path,
3898 len2, p->two->path,
3899 len1, p->one->path,
3900 len2, p->two->path);
3901 git_SHA1_Update(&ctx, buffer, len1);
3902
3903 if (diff_filespec_is_binary(p->one) ||
3904 diff_filespec_is_binary(p->two)) {
3905 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3906 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3907 continue;
3908 }
3909
3910 xpp.flags = 0;
3911 xecfg.ctxlen = 3;
3912 xecfg.flags = 0;
3913 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3914 &xpp, &xecfg);
3915 }
3916
3917 git_SHA1_Final(sha1, &ctx);
3918 return 0;
3919 }
3920
3921 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3922 {
3923 struct diff_queue_struct *q = &diff_queued_diff;
3924 int i;
3925 int result = diff_get_patch_id(options, sha1);
3926
3927 for (i = 0; i < q->nr; i++)
3928 diff_free_filepair(q->queue[i]);
3929
3930 free(q->queue);
3931 DIFF_QUEUE_CLEAR(q);
3932
3933 return result;
3934 }
3935
3936 static int is_summary_empty(const struct diff_queue_struct *q)
3937 {
3938 int i;
3939
3940 for (i = 0; i < q->nr; i++) {
3941 const struct diff_filepair *p = q->queue[i];
3942
3943 switch (p->status) {
3944 case DIFF_STATUS_DELETED:
3945 case DIFF_STATUS_ADDED:
3946 case DIFF_STATUS_COPIED:
3947 case DIFF_STATUS_RENAMED:
3948 return 0;
3949 default:
3950 if (p->score)
3951 return 0;
3952 if (p->one->mode && p->two->mode &&
3953 p->one->mode != p->two->mode)
3954 return 0;
3955 break;
3956 }
3957 }
3958 return 1;
3959 }
3960
3961 void diff_flush(struct diff_options *options)
3962 {
3963 struct diff_queue_struct *q = &diff_queued_diff;
3964 int i, output_format = options->output_format;
3965 int separator = 0;
3966
3967 /*
3968 * Order: raw, stat, summary, patch
3969 * or: name/name-status/checkdiff (other bits clear)
3970 */
3971 if (!q->nr)
3972 goto free_queue;
3973
3974 if (output_format & (DIFF_FORMAT_RAW |
3975 DIFF_FORMAT_NAME |
3976 DIFF_FORMAT_NAME_STATUS |
3977 DIFF_FORMAT_CHECKDIFF)) {
3978 for (i = 0; i < q->nr; i++) {
3979 struct diff_filepair *p = q->queue[i];
3980 if (check_pair_status(p))
3981 flush_one_pair(p, options);
3982 }
3983 separator++;
3984 }
3985
3986 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
3987 struct diffstat_t diffstat;
3988
3989 memset(&diffstat, 0, sizeof(struct diffstat_t));
3990 for (i = 0; i < q->nr; i++) {
3991 struct diff_filepair *p = q->queue[i];
3992 if (check_pair_status(p))
3993 diff_flush_stat(p, options, &diffstat);
3994 }
3995 if (output_format & DIFF_FORMAT_NUMSTAT)
3996 show_numstat(&diffstat, options);
3997 if (output_format & DIFF_FORMAT_DIFFSTAT)
3998 show_stats(&diffstat, options);
3999 if (output_format & DIFF_FORMAT_SHORTSTAT)
4000 show_shortstats(&diffstat, options);
4001 free_diffstat_info(&diffstat);
4002 separator++;
4003 }
4004 if (output_format & DIFF_FORMAT_DIRSTAT)
4005 show_dirstat(options);
4006
4007 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4008 for (i = 0; i < q->nr; i++) {
4009 diff_summary(options, q->queue[i]);
4010 }
4011 separator++;
4012 }
4013
4014 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4015 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4016 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4017 /*
4018 * run diff_flush_patch for the exit status. setting
4019 * options->file to /dev/null should be safe, becaue we
4020 * aren't supposed to produce any output anyway.
4021 */
4022 if (options->close_file)
4023 fclose(options->file);
4024 options->file = fopen("/dev/null", "w");
4025 if (!options->file)
4026 die_errno("Could not open /dev/null");
4027 options->close_file = 1;
4028 for (i = 0; i < q->nr; i++) {
4029 struct diff_filepair *p = q->queue[i];
4030 if (check_pair_status(p))
4031 diff_flush_patch(p, options);
4032 if (options->found_changes)
4033 break;
4034 }
4035 }
4036
4037 if (output_format & DIFF_FORMAT_PATCH) {
4038 if (separator) {
4039 putc(options->line_termination, options->file);
4040 if (options->stat_sep) {
4041 /* attach patch instead of inline */
4042 fputs(options->stat_sep, options->file);
4043 }
4044 }
4045
4046 for (i = 0; i < q->nr; i++) {
4047 struct diff_filepair *p = q->queue[i];
4048 if (check_pair_status(p))
4049 diff_flush_patch(p, options);
4050 }
4051 }
4052
4053 if (output_format & DIFF_FORMAT_CALLBACK)
4054 options->format_callback(q, options, options->format_callback_data);
4055
4056 for (i = 0; i < q->nr; i++)
4057 diff_free_filepair(q->queue[i]);
4058 free_queue:
4059 free(q->queue);
4060 DIFF_QUEUE_CLEAR(q);
4061 if (options->close_file)
4062 fclose(options->file);
4063
4064 /*
4065 * Report the content-level differences with HAS_CHANGES;
4066 * diff_addremove/diff_change does not set the bit when
4067 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4068 */
4069 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4070 if (options->found_changes)
4071 DIFF_OPT_SET(options, HAS_CHANGES);
4072 else
4073 DIFF_OPT_CLR(options, HAS_CHANGES);
4074 }
4075 }
4076
4077 static void diffcore_apply_filter(const char *filter)
4078 {
4079 int i;
4080 struct diff_queue_struct *q = &diff_queued_diff;
4081 struct diff_queue_struct outq;
4082 DIFF_QUEUE_CLEAR(&outq);
4083
4084 if (!filter)
4085 return;
4086
4087 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4088 int found;
4089 for (i = found = 0; !found && i < q->nr; i++) {
4090 struct diff_filepair *p = q->queue[i];
4091 if (((p->status == DIFF_STATUS_MODIFIED) &&
4092 ((p->score &&
4093 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4094 (!p->score &&
4095 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4096 ((p->status != DIFF_STATUS_MODIFIED) &&
4097 strchr(filter, p->status)))
4098 found++;
4099 }
4100 if (found)
4101 return;
4102
4103 /* otherwise we will clear the whole queue
4104 * by copying the empty outq at the end of this
4105 * function, but first clear the current entries
4106 * in the queue.
4107 */
4108 for (i = 0; i < q->nr; i++)
4109 diff_free_filepair(q->queue[i]);
4110 }
4111 else {
4112 /* Only the matching ones */
4113 for (i = 0; i < q->nr; i++) {
4114 struct diff_filepair *p = q->queue[i];
4115
4116 if (((p->status == DIFF_STATUS_MODIFIED) &&
4117 ((p->score &&
4118 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4119 (!p->score &&
4120 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4121 ((p->status != DIFF_STATUS_MODIFIED) &&
4122 strchr(filter, p->status)))
4123 diff_q(&outq, p);
4124 else
4125 diff_free_filepair(p);
4126 }
4127 }
4128 free(q->queue);
4129 *q = outq;
4130 }
4131
4132 /* Check whether two filespecs with the same mode and size are identical */
4133 static int diff_filespec_is_identical(struct diff_filespec *one,
4134 struct diff_filespec *two)
4135 {
4136 if (S_ISGITLINK(one->mode))
4137 return 0;
4138 if (diff_populate_filespec(one, 0))
4139 return 0;
4140 if (diff_populate_filespec(two, 0))
4141 return 0;
4142 return !memcmp(one->data, two->data, one->size);
4143 }
4144
4145 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4146 {
4147 int i;
4148 struct diff_queue_struct *q = &diff_queued_diff;
4149 struct diff_queue_struct outq;
4150 DIFF_QUEUE_CLEAR(&outq);
4151
4152 for (i = 0; i < q->nr; i++) {
4153 struct diff_filepair *p = q->queue[i];
4154
4155 /*
4156 * 1. Entries that come from stat info dirtiness
4157 * always have both sides (iow, not create/delete),
4158 * one side of the object name is unknown, with
4159 * the same mode and size. Keep the ones that
4160 * do not match these criteria. They have real
4161 * differences.
4162 *
4163 * 2. At this point, the file is known to be modified,
4164 * with the same mode and size, and the object
4165 * name of one side is unknown. Need to inspect
4166 * the identical contents.
4167 */
4168 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4169 !DIFF_FILE_VALID(p->two) ||
4170 (p->one->sha1_valid && p->two->sha1_valid) ||
4171 (p->one->mode != p->two->mode) ||
4172 diff_populate_filespec(p->one, 1) ||
4173 diff_populate_filespec(p->two, 1) ||
4174 (p->one->size != p->two->size) ||
4175 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4176 diff_q(&outq, p);
4177 else {
4178 /*
4179 * The caller can subtract 1 from skip_stat_unmatch
4180 * to determine how many paths were dirty only
4181 * due to stat info mismatch.
4182 */
4183 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4184 diffopt->skip_stat_unmatch++;
4185 diff_free_filepair(p);
4186 }
4187 }
4188 free(q->queue);
4189 *q = outq;
4190 }
4191
4192 static int diffnamecmp(const void *a_, const void *b_)
4193 {
4194 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4195 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4196 const char *name_a, *name_b;
4197
4198 name_a = a->one ? a->one->path : a->two->path;
4199 name_b = b->one ? b->one->path : b->two->path;
4200 return strcmp(name_a, name_b);
4201 }
4202
4203 void diffcore_fix_diff_index(struct diff_options *options)
4204 {
4205 struct diff_queue_struct *q = &diff_queued_diff;
4206 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4207 }
4208
4209 void diffcore_std(struct diff_options *options)
4210 {
4211 if (options->skip_stat_unmatch)
4212 diffcore_skip_stat_unmatch(options);
4213 if (!options->found_follow) {
4214 /* See try_to_follow_renames() in tree-diff.c */
4215 if (options->break_opt != -1)
4216 diffcore_break(options->break_opt);
4217 if (options->detect_rename)
4218 diffcore_rename(options);
4219 if (options->break_opt != -1)
4220 diffcore_merge_broken();
4221 }
4222 if (options->pickaxe)
4223 diffcore_pickaxe(options);
4224 if (options->orderfile)
4225 diffcore_order(options->orderfile);
4226 if (!options->found_follow)
4227 /* See try_to_follow_renames() in tree-diff.c */
4228 diff_resolve_rename_copy();
4229 diffcore_apply_filter(options->filter);
4230
4231 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4232 DIFF_OPT_SET(options, HAS_CHANGES);
4233 else
4234 DIFF_OPT_CLR(options, HAS_CHANGES);
4235
4236 options->found_follow = 0;
4237 }
4238
4239 int diff_result_code(struct diff_options *opt, int status)
4240 {
4241 int result = 0;
4242 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4243 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4244 return status;
4245 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4246 DIFF_OPT_TST(opt, HAS_CHANGES))
4247 result |= 01;
4248 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4249 DIFF_OPT_TST(opt, CHECK_FAILED))
4250 result |= 02;
4251 return result;
4252 }
4253
4254 /*
4255 * Shall changes to this submodule be ignored?
4256 *
4257 * Submodule changes can be configured to be ignored separately for each path,
4258 * but that configuration can be overridden from the command line.
4259 */
4260 static int is_submodule_ignored(const char *path, struct diff_options *options)
4261 {
4262 int ignored = 0;
4263 unsigned orig_flags = options->flags;
4264 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4265 set_diffopt_flags_from_submodule_config(options, path);
4266 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4267 ignored = 1;
4268 options->flags = orig_flags;
4269 return ignored;
4270 }
4271
4272 void diff_addremove(struct diff_options *options,
4273 int addremove, unsigned mode,
4274 const unsigned char *sha1,
4275 const char *concatpath, unsigned dirty_submodule)
4276 {
4277 struct diff_filespec *one, *two;
4278
4279 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4280 return;
4281
4282 /* This may look odd, but it is a preparation for
4283 * feeding "there are unchanged files which should
4284 * not produce diffs, but when you are doing copy
4285 * detection you would need them, so here they are"
4286 * entries to the diff-core. They will be prefixed
4287 * with something like '=' or '*' (I haven't decided
4288 * which but should not make any difference).
4289 * Feeding the same new and old to diff_change()
4290 * also has the same effect.
4291 * Before the final output happens, they are pruned after
4292 * merged into rename/copy pairs as appropriate.
4293 */
4294 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4295 addremove = (addremove == '+' ? '-' :
4296 addremove == '-' ? '+' : addremove);
4297
4298 if (options->prefix &&
4299 strncmp(concatpath, options->prefix, options->prefix_length))
4300 return;
4301
4302 one = alloc_filespec(concatpath);
4303 two = alloc_filespec(concatpath);
4304
4305 if (addremove != '+')
4306 fill_filespec(one, sha1, mode);
4307 if (addremove != '-') {
4308 fill_filespec(two, sha1, mode);
4309 two->dirty_submodule = dirty_submodule;
4310 }
4311
4312 diff_queue(&diff_queued_diff, one, two);
4313 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4314 DIFF_OPT_SET(options, HAS_CHANGES);
4315 }
4316
4317 void diff_change(struct diff_options *options,
4318 unsigned old_mode, unsigned new_mode,
4319 const unsigned char *old_sha1,
4320 const unsigned char *new_sha1,
4321 const char *concatpath,
4322 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4323 {
4324 struct diff_filespec *one, *two;
4325
4326 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4327 is_submodule_ignored(concatpath, options))
4328 return;
4329
4330 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4331 unsigned tmp;
4332 const unsigned char *tmp_c;
4333 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4334 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4335 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4336 new_dirty_submodule = tmp;
4337 }
4338
4339 if (options->prefix &&
4340 strncmp(concatpath, options->prefix, options->prefix_length))
4341 return;
4342
4343 one = alloc_filespec(concatpath);
4344 two = alloc_filespec(concatpath);
4345 fill_filespec(one, old_sha1, old_mode);
4346 fill_filespec(two, new_sha1, new_mode);
4347 one->dirty_submodule = old_dirty_submodule;
4348 two->dirty_submodule = new_dirty_submodule;
4349
4350 diff_queue(&diff_queued_diff, one, two);
4351 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4352 DIFF_OPT_SET(options, HAS_CHANGES);
4353 }
4354
4355 void diff_unmerge(struct diff_options *options,
4356 const char *path,
4357 unsigned mode, const unsigned char *sha1)
4358 {
4359 struct diff_filespec *one, *two;
4360
4361 if (options->prefix &&
4362 strncmp(path, options->prefix, options->prefix_length))
4363 return;
4364
4365 one = alloc_filespec(path);
4366 two = alloc_filespec(path);
4367 fill_filespec(one, sha1, mode);
4368 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4369 }
4370
4371 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4372 size_t *outsize)
4373 {
4374 struct diff_tempfile *temp;
4375 const char *argv[3];
4376 const char **arg = argv;
4377 struct child_process child;
4378 struct strbuf buf = STRBUF_INIT;
4379 int err = 0;
4380
4381 temp = prepare_temp_file(spec->path, spec);
4382 *arg++ = pgm;
4383 *arg++ = temp->name;
4384 *arg = NULL;
4385
4386 memset(&child, 0, sizeof(child));
4387 child.use_shell = 1;
4388 child.argv = argv;
4389 child.out = -1;
4390 if (start_command(&child)) {
4391 remove_tempfile();
4392 return NULL;
4393 }
4394
4395 if (strbuf_read(&buf, child.out, 0) < 0)
4396 err = error("error reading from textconv command '%s'", pgm);
4397 close(child.out);
4398
4399 if (finish_command(&child) || err) {
4400 strbuf_release(&buf);
4401 remove_tempfile();
4402 return NULL;
4403 }
4404 remove_tempfile();
4405
4406 return strbuf_detach(&buf, outsize);
4407 }
4408
4409 size_t fill_textconv(struct userdiff_driver *driver,
4410 struct diff_filespec *df,
4411 char **outbuf)
4412 {
4413 size_t size;
4414
4415 if (!driver || !driver->textconv) {
4416 if (!DIFF_FILE_VALID(df)) {
4417 *outbuf = "";
4418 return 0;
4419 }
4420 if (diff_populate_filespec(df, 0))
4421 die("unable to read files to diff");
4422 *outbuf = df->data;
4423 return df->size;
4424 }
4425
4426 if (driver->textconv_cache && df->sha1_valid) {
4427 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4428 &size);
4429 if (*outbuf)
4430 return size;
4431 }
4432
4433 *outbuf = run_textconv(driver->textconv, df, &size);
4434 if (!*outbuf)
4435 die("unable to read files to diff");
4436
4437 if (driver->textconv_cache && df->sha1_valid) {
4438 /* ignore errors, as we might be in a readonly repository */
4439 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4440 size);
4441 /*
4442 * we could save up changes and flush them all at the end,
4443 * but we would need an extra call after all diffing is done.
4444 * Since generating a cache entry is the slow path anyway,
4445 * this extra overhead probably isn't a big deal.
4446 */
4447 notes_cache_write(driver->textconv_cache);
4448 }
4449
4450 return size;
4451 }
Michael J Gruber's git tree