?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Document the underlying protocol used by shallow repositories and --depth commands.
[git/jnareb-git.git] / diff.c
blob559bf574a88f97402f32e1bd4ab51f6f0c9d5026
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 400;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static struct diff_options default_diff_options;
35
36 static char diff_colors[][COLOR_MAXLEN] = {
37 GIT_COLOR_RESET,
38 GIT_COLOR_NORMAL, /* PLAIN */
39 GIT_COLOR_BOLD, /* METAINFO */
40 GIT_COLOR_CYAN, /* FRAGINFO */
41 GIT_COLOR_RED, /* OLD */
42 GIT_COLOR_GREEN, /* NEW */
43 GIT_COLOR_YELLOW, /* COMMIT */
44 GIT_COLOR_BG_RED, /* WHITESPACE */
45 GIT_COLOR_NORMAL, /* FUNCINFO */
46 };
47
48 static int parse_diff_color_slot(const char *var, int ofs)
49 {
50 if (!strcasecmp(var+ofs, "plain"))
51 return DIFF_PLAIN;
52 if (!strcasecmp(var+ofs, "meta"))
53 return DIFF_METAINFO;
54 if (!strcasecmp(var+ofs, "frag"))
55 return DIFF_FRAGINFO;
56 if (!strcasecmp(var+ofs, "old"))
57 return DIFF_FILE_OLD;
58 if (!strcasecmp(var+ofs, "new"))
59 return DIFF_FILE_NEW;
60 if (!strcasecmp(var+ofs, "commit"))
61 return DIFF_COMMIT;
62 if (!strcasecmp(var+ofs, "whitespace"))
63 return DIFF_WHITESPACE;
64 if (!strcasecmp(var+ofs, "func"))
65 return DIFF_FUNCINFO;
66 return -1;
67 }
68
69 static int git_config_rename(const char *var, const char *value)
70 {
71 if (!value)
72 return DIFF_DETECT_RENAME;
73 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
74 return DIFF_DETECT_COPY;
75 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
76 }
77
78 /*
79 * These are to give UI layer defaults.
80 * The core-level commands such as git-diff-files should
81 * never be affected by the setting of diff.renames
82 * the user happens to have in the configuration file.
83 */
84 int git_diff_ui_config(const char *var, const char *value, void *cb)
85 {
86 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
87 diff_use_color_default = git_config_colorbool(var, value, -1);
88 return 0;
89 }
90 if (!strcmp(var, "diff.renames")) {
91 diff_detect_rename_default = git_config_rename(var, value);
92 return 0;
93 }
94 if (!strcmp(var, "diff.autorefreshindex")) {
95 diff_auto_refresh_index = git_config_bool(var, value);
96 return 0;
97 }
98 if (!strcmp(var, "diff.mnemonicprefix")) {
99 diff_mnemonic_prefix = git_config_bool(var, value);
100 return 0;
101 }
102 if (!strcmp(var, "diff.noprefix")) {
103 diff_no_prefix = git_config_bool(var, value);
104 return 0;
105 }
106 if (!strcmp(var, "diff.external"))
107 return git_config_string(&external_diff_cmd_cfg, var, value);
108 if (!strcmp(var, "diff.wordregex"))
109 return git_config_string(&diff_word_regex_cfg, var, value);
110
111 if (!strcmp(var, "diff.ignoresubmodules"))
112 handle_ignore_submodules_arg(&default_diff_options, value);
113
114 return git_diff_basic_config(var, value, cb);
115 }
116
117 int git_diff_basic_config(const char *var, const char *value, void *cb)
118 {
119 if (!strcmp(var, "diff.renamelimit")) {
120 diff_rename_limit_default = git_config_int(var, value);
121 return 0;
122 }
123
124 switch (userdiff_config(var, value)) {
125 case 0: break;
126 case -1: return -1;
127 default: return 0;
128 }
129
130 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
131 int slot = parse_diff_color_slot(var, 11);
132 if (slot < 0)
133 return 0;
134 if (!value)
135 return config_error_nonbool(var);
136 color_parse(value, var, diff_colors[slot]);
137 return 0;
138 }
139
140 /* like GNU diff's --suppress-blank-empty option */
141 if (!strcmp(var, "diff.suppressblankempty") ||
142 /* for backwards compatibility */
143 !strcmp(var, "diff.suppress-blank-empty")) {
144 diff_suppress_blank_empty = git_config_bool(var, value);
145 return 0;
146 }
147
148 if (!prefixcmp(var, "submodule."))
149 return parse_submodule_config_option(var, value);
150
151 return git_color_default_config(var, value, cb);
152 }
153
154 static char *quote_two(const char *one, const char *two)
155 {
156 int need_one = quote_c_style(one, NULL, NULL, 1);
157 int need_two = quote_c_style(two, NULL, NULL, 1);
158 struct strbuf res = STRBUF_INIT;
159
160 if (need_one + need_two) {
161 strbuf_addch(&res, '"');
162 quote_c_style(one, &res, NULL, 1);
163 quote_c_style(two, &res, NULL, 1);
164 strbuf_addch(&res, '"');
165 } else {
166 strbuf_addstr(&res, one);
167 strbuf_addstr(&res, two);
168 }
169 return strbuf_detach(&res, NULL);
170 }
171
172 static const char *external_diff(void)
173 {
174 static const char *external_diff_cmd = NULL;
175 static int done_preparing = 0;
176
177 if (done_preparing)
178 return external_diff_cmd;
179 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
180 if (!external_diff_cmd)
181 external_diff_cmd = external_diff_cmd_cfg;
182 done_preparing = 1;
183 return external_diff_cmd;
184 }
185
186 static struct diff_tempfile {
187 const char *name; /* filename external diff should read from */
188 char hex[41];
189 char mode[10];
190 char tmp_path[PATH_MAX];
191 } diff_temp[2];
192
193 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
194
195 struct emit_callback {
196 int color_diff;
197 unsigned ws_rule;
198 int blank_at_eof_in_preimage;
199 int blank_at_eof_in_postimage;
200 int lno_in_preimage;
201 int lno_in_postimage;
202 sane_truncate_fn truncate;
203 const char **label_path;
204 struct diff_words_data *diff_words;
205 struct diff_options *opt;
206 int *found_changesp;
207 struct strbuf *header;
208 };
209
210 static int count_lines(const char *data, int size)
211 {
212 int count, ch, completely_empty = 1, nl_just_seen = 0;
213 count = 0;
214 while (0 < size--) {
215 ch = *data++;
216 if (ch == '\n') {
217 count++;
218 nl_just_seen = 1;
219 completely_empty = 0;
220 }
221 else {
222 nl_just_seen = 0;
223 completely_empty = 0;
224 }
225 }
226 if (completely_empty)
227 return 0;
228 if (!nl_just_seen)
229 count++; /* no trailing newline */
230 return count;
231 }
232
233 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
234 {
235 if (!DIFF_FILE_VALID(one)) {
236 mf->ptr = (char *)""; /* does not matter */
237 mf->size = 0;
238 return 0;
239 }
240 else if (diff_populate_filespec(one, 0))
241 return -1;
242
243 mf->ptr = one->data;
244 mf->size = one->size;
245 return 0;
246 }
247
248 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
249 static unsigned long diff_filespec_size(struct diff_filespec *one)
250 {
251 if (!DIFF_FILE_VALID(one))
252 return 0;
253 diff_populate_filespec(one, 1);
254 return one->size;
255 }
256
257 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
258 {
259 char *ptr = mf->ptr;
260 long size = mf->size;
261 int cnt = 0;
262
263 if (!size)
264 return cnt;
265 ptr += size - 1; /* pointing at the very end */
266 if (*ptr != '\n')
267 ; /* incomplete line */
268 else
269 ptr--; /* skip the last LF */
270 while (mf->ptr < ptr) {
271 char *prev_eol;
272 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
273 if (*prev_eol == '\n')
274 break;
275 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
276 break;
277 cnt++;
278 ptr = prev_eol - 1;
279 }
280 return cnt;
281 }
282
283 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
284 struct emit_callback *ecbdata)
285 {
286 int l1, l2, at;
287 unsigned ws_rule = ecbdata->ws_rule;
288 l1 = count_trailing_blank(mf1, ws_rule);
289 l2 = count_trailing_blank(mf2, ws_rule);
290 if (l2 <= l1) {
291 ecbdata->blank_at_eof_in_preimage = 0;
292 ecbdata->blank_at_eof_in_postimage = 0;
293 return;
294 }
295 at = count_lines(mf1->ptr, mf1->size);
296 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
297
298 at = count_lines(mf2->ptr, mf2->size);
299 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
300 }
301
302 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
303 int first, const char *line, int len)
304 {
305 int has_trailing_newline, has_trailing_carriage_return;
306 int nofirst;
307 FILE *file = o->file;
308
309 if (o->output_prefix) {
310 struct strbuf *msg = NULL;
311 msg = o->output_prefix(o, o->output_prefix_data);
312 assert(msg);
313 fwrite(msg->buf, msg->len, 1, file);
314 }
315
316 if (len == 0) {
317 has_trailing_newline = (first == '\n');
318 has_trailing_carriage_return = (!has_trailing_newline &&
319 (first == '\r'));
320 nofirst = has_trailing_newline || has_trailing_carriage_return;
321 } else {
322 has_trailing_newline = (len > 0 && line[len-1] == '\n');
323 if (has_trailing_newline)
324 len--;
325 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
326 if (has_trailing_carriage_return)
327 len--;
328 nofirst = 0;
329 }
330
331 if (len || !nofirst) {
332 fputs(set, file);
333 if (!nofirst)
334 fputc(first, file);
335 fwrite(line, len, 1, file);
336 fputs(reset, file);
337 }
338 if (has_trailing_carriage_return)
339 fputc('\r', file);
340 if (has_trailing_newline)
341 fputc('\n', file);
342 }
343
344 static void emit_line(struct diff_options *o, const char *set, const char *reset,
345 const char *line, int len)
346 {
347 emit_line_0(o, set, reset, line[0], line+1, len-1);
348 }
349
350 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
351 {
352 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
353 ecbdata->blank_at_eof_in_preimage &&
354 ecbdata->blank_at_eof_in_postimage &&
355 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
356 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
357 return 0;
358 return ws_blank_line(line, len, ecbdata->ws_rule);
359 }
360
361 static void emit_add_line(const char *reset,
362 struct emit_callback *ecbdata,
363 const char *line, int len)
364 {
365 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
366 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
367
368 if (!*ws)
369 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
370 else if (new_blank_line_at_eof(ecbdata, line, len))
371 /* Blank line at EOF - paint '+' as well */
372 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
373 else {
374 /* Emit just the prefix, then the rest. */
375 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
376 ws_check_emit(line, len, ecbdata->ws_rule,
377 ecbdata->opt->file, set, reset, ws);
378 }
379 }
380
381 static void emit_hunk_header(struct emit_callback *ecbdata,
382 const char *line, int len)
383 {
384 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
385 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
386 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
387 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
388 static const char atat[2] = { '@', '@' };
389 const char *cp, *ep;
390 struct strbuf msgbuf = STRBUF_INIT;
391 int org_len = len;
392 int i = 1;
393
394 /*
395 * As a hunk header must begin with "@@ -<old>, +<new> @@",
396 * it always is at least 10 bytes long.
397 */
398 if (len < 10 ||
399 memcmp(line, atat, 2) ||
400 !(ep = memmem(line + 2, len - 2, atat, 2))) {
401 emit_line(ecbdata->opt, plain, reset, line, len);
402 return;
403 }
404 ep += 2; /* skip over @@ */
405
406 /* The hunk header in fraginfo color */
407 strbuf_add(&msgbuf, frag, strlen(frag));
408 strbuf_add(&msgbuf, line, ep - line);
409 strbuf_add(&msgbuf, reset, strlen(reset));
410
411 /*
412 * trailing "\r\n"
413 */
414 for ( ; i < 3; i++)
415 if (line[len - i] == '\r' || line[len - i] == '\n')
416 len--;
417
418 /* blank before the func header */
419 for (cp = ep; ep - line < len; ep++)
420 if (*ep != ' ' && *ep != '\t')
421 break;
422 if (ep != cp) {
423 strbuf_add(&msgbuf, plain, strlen(plain));
424 strbuf_add(&msgbuf, cp, ep - cp);
425 strbuf_add(&msgbuf, reset, strlen(reset));
426 }
427
428 if (ep < line + len) {
429 strbuf_add(&msgbuf, func, strlen(func));
430 strbuf_add(&msgbuf, ep, line + len - ep);
431 strbuf_add(&msgbuf, reset, strlen(reset));
432 }
433
434 strbuf_add(&msgbuf, line + len, org_len - len);
435 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
436 strbuf_release(&msgbuf);
437 }
438
439 static struct diff_tempfile *claim_diff_tempfile(void) {
440 int i;
441 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
442 if (!diff_temp[i].name)
443 return diff_temp + i;
444 die("BUG: diff is failing to clean up its tempfiles");
445 }
446
447 static int remove_tempfile_installed;
448
449 static void remove_tempfile(void)
450 {
451 int i;
452 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
453 if (diff_temp[i].name == diff_temp[i].tmp_path)
454 unlink_or_warn(diff_temp[i].name);
455 diff_temp[i].name = NULL;
456 }
457 }
458
459 static void remove_tempfile_on_signal(int signo)
460 {
461 remove_tempfile();
462 sigchain_pop(signo);
463 raise(signo);
464 }
465
466 static void print_line_count(FILE *file, int count)
467 {
468 switch (count) {
469 case 0:
470 fprintf(file, "0,0");
471 break;
472 case 1:
473 fprintf(file, "1");
474 break;
475 default:
476 fprintf(file, "1,%d", count);
477 break;
478 }
479 }
480
481 static void emit_rewrite_lines(struct emit_callback *ecb,
482 int prefix, const char *data, int size)
483 {
484 const char *endp = NULL;
485 static const char *nneof = " No newline at end of file\n";
486 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
487 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
488
489 while (0 < size) {
490 int len;
491
492 endp = memchr(data, '\n', size);
493 len = endp ? (endp - data + 1) : size;
494 if (prefix != '+') {
495 ecb->lno_in_preimage++;
496 emit_line_0(ecb->opt, old, reset, '-',
497 data, len);
498 } else {
499 ecb->lno_in_postimage++;
500 emit_add_line(reset, ecb, data, len);
501 }
502 size -= len;
503 data += len;
504 }
505 if (!endp) {
506 const char *plain = diff_get_color(ecb->color_diff,
507 DIFF_PLAIN);
508 emit_line_0(ecb->opt, plain, reset, '\\',
509 nneof, strlen(nneof));
510 }
511 }
512
513 static void emit_rewrite_diff(const char *name_a,
514 const char *name_b,
515 struct diff_filespec *one,
516 struct diff_filespec *two,
517 struct userdiff_driver *textconv_one,
518 struct userdiff_driver *textconv_two,
519 struct diff_options *o)
520 {
521 int lc_a, lc_b;
522 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
523 const char *name_a_tab, *name_b_tab;
524 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
525 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
526 const char *reset = diff_get_color(color_diff, DIFF_RESET);
527 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
528 const char *a_prefix, *b_prefix;
529 char *data_one, *data_two;
530 size_t size_one, size_two;
531 struct emit_callback ecbdata;
532 char *line_prefix = "";
533 struct strbuf *msgbuf;
534
535 if (o && o->output_prefix) {
536 msgbuf = o->output_prefix(o, o->output_prefix_data);
537 line_prefix = msgbuf->buf;
538 }
539
540 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
541 a_prefix = o->b_prefix;
542 b_prefix = o->a_prefix;
543 } else {
544 a_prefix = o->a_prefix;
545 b_prefix = o->b_prefix;
546 }
547
548 name_a += (*name_a == '/');
549 name_b += (*name_b == '/');
550 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
551 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
552
553 strbuf_reset(&a_name);
554 strbuf_reset(&b_name);
555 quote_two_c_style(&a_name, a_prefix, name_a, 0);
556 quote_two_c_style(&b_name, b_prefix, name_b, 0);
557
558 size_one = fill_textconv(textconv_one, one, &data_one);
559 size_two = fill_textconv(textconv_two, two, &data_two);
560
561 memset(&ecbdata, 0, sizeof(ecbdata));
562 ecbdata.color_diff = color_diff;
563 ecbdata.found_changesp = &o->found_changes;
564 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
565 ecbdata.opt = o;
566 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
567 mmfile_t mf1, mf2;
568 mf1.ptr = (char *)data_one;
569 mf2.ptr = (char *)data_two;
570 mf1.size = size_one;
571 mf2.size = size_two;
572 check_blank_at_eof(&mf1, &mf2, &ecbdata);
573 }
574 ecbdata.lno_in_preimage = 1;
575 ecbdata.lno_in_postimage = 1;
576
577 lc_a = count_lines(data_one, size_one);
578 lc_b = count_lines(data_two, size_two);
579 fprintf(o->file,
580 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
581 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
582 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
583 line_prefix, fraginfo);
584 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 const char *prefix;
620 const char *suffix;
621 const char *color; /* NULL; filled in by the setup code if
622 * color is enabled */
623 };
624
625 struct diff_words_style {
626 enum diff_words_type type;
627 struct diff_words_style_elem new, old, ctx;
628 const char *newline;
629 };
630
631 static struct diff_words_style diff_words_styles[] = {
632 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
633 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
634 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
635 };
636
637 struct diff_words_data {
638 struct diff_words_buffer minus, plus;
639 const char *current_plus;
640 int last_minus;
641 struct diff_options *opt;
642 regex_t *word_regex;
643 enum diff_words_type type;
644 struct diff_words_style *style;
645 };
646
647 static int fn_out_diff_words_write_helper(FILE *fp,
648 struct diff_words_style_elem *st_el,
649 const char *newline,
650 size_t count, const char *buf,
651 const char *line_prefix)
652 {
653 int print = 0;
654
655 while (count) {
656 char *p = memchr(buf, '\n', count);
657 if (print)
658 fputs(line_prefix, fp);
659 if (p != buf) {
660 if (st_el->color && fputs(st_el->color, fp) < 0)
661 return -1;
662 if (fputs(st_el->prefix, fp) < 0 ||
663 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
664 fputs(st_el->suffix, fp) < 0)
665 return -1;
666 if (st_el->color && *st_el->color
667 && fputs(GIT_COLOR_RESET, fp) < 0)
668 return -1;
669 }
670 if (!p)
671 return 0;
672 if (fputs(newline, fp) < 0)
673 return -1;
674 count -= p + 1 - buf;
675 buf = p + 1;
676 print = 1;
677 }
678 return 0;
679 }
680
681 /*
682 * '--color-words' algorithm can be described as:
683 *
684 * 1. collect a the minus/plus lines of a diff hunk, divided into
685 * minus-lines and plus-lines;
686 *
687 * 2. break both minus-lines and plus-lines into words and
688 * place them into two mmfile_t with one word for each line;
689 *
690 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
691 *
692 * And for the common parts of the both file, we output the plus side text.
693 * diff_words->current_plus is used to trace the current position of the plus file
694 * which printed. diff_words->last_minus is used to trace the last minus word
695 * printed.
696 *
697 * For '--graph' to work with '--color-words', we need to output the graph prefix
698 * on each line of color words output. Generally, there are two conditions on
699 * which we should output the prefix.
700 *
701 * 1. diff_words->last_minus == 0 &&
702 * diff_words->current_plus == diff_words->plus.text.ptr
703 *
704 * that is: the plus text must start as a new line, and if there is no minus
705 * word printed, a graph prefix must be printed.
706 *
707 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
708 * *(diff_words->current_plus - 1) == '\n'
709 *
710 * that is: a graph prefix must be printed following a '\n'
711 */
712 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
713 {
714 if ((diff_words->last_minus == 0 &&
715 diff_words->current_plus == diff_words->plus.text.ptr) ||
716 (diff_words->current_plus > diff_words->plus.text.ptr &&
717 *(diff_words->current_plus - 1) == '\n')) {
718 return 1;
719 } else {
720 return 0;
721 }
722 }
723
724 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
725 {
726 struct diff_words_data *diff_words = priv;
727 struct diff_words_style *style = diff_words->style;
728 int minus_first, minus_len, plus_first, plus_len;
729 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
730 struct diff_options *opt = diff_words->opt;
731 struct strbuf *msgbuf;
732 char *line_prefix = "";
733
734 if (line[0] != '@' || parse_hunk_header(line, len,
735 &minus_first, &minus_len, &plus_first, &plus_len))
736 return;
737
738 assert(opt);
739 if (opt->output_prefix) {
740 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
741 line_prefix = msgbuf->buf;
742 }
743
744 /* POSIX requires that first be decremented by one if len == 0... */
745 if (minus_len) {
746 minus_begin = diff_words->minus.orig[minus_first].begin;
747 minus_end =
748 diff_words->minus.orig[minus_first + minus_len - 1].end;
749 } else
750 minus_begin = minus_end =
751 diff_words->minus.orig[minus_first].end;
752
753 if (plus_len) {
754 plus_begin = diff_words->plus.orig[plus_first].begin;
755 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
756 } else
757 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
758
759 if (color_words_output_graph_prefix(diff_words)) {
760 fputs(line_prefix, diff_words->opt->file);
761 }
762 if (diff_words->current_plus != plus_begin) {
763 fn_out_diff_words_write_helper(diff_words->opt->file,
764 &style->ctx, style->newline,
765 plus_begin - diff_words->current_plus,
766 diff_words->current_plus, line_prefix);
767 if (*(plus_begin - 1) == '\n')
768 fputs(line_prefix, diff_words->opt->file);
769 }
770 if (minus_begin != minus_end) {
771 fn_out_diff_words_write_helper(diff_words->opt->file,
772 &style->old, style->newline,
773 minus_end - minus_begin, minus_begin,
774 line_prefix);
775 }
776 if (plus_begin != plus_end) {
777 fn_out_diff_words_write_helper(diff_words->opt->file,
778 &style->new, style->newline,
779 plus_end - plus_begin, plus_begin,
780 line_prefix);
781 }
782
783 diff_words->current_plus = plus_end;
784 diff_words->last_minus = minus_first;
785 }
786
787 /* This function starts looking at *begin, and returns 0 iff a word was found. */
788 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
789 int *begin, int *end)
790 {
791 if (word_regex && *begin < buffer->size) {
792 regmatch_t match[1];
793 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
794 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
795 '\n', match[0].rm_eo - match[0].rm_so);
796 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
797 *begin += match[0].rm_so;
798 return *begin >= *end;
799 }
800 return -1;
801 }
802
803 /* find the next word */
804 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
805 (*begin)++;
806 if (*begin >= buffer->size)
807 return -1;
808
809 /* find the end of the word */
810 *end = *begin + 1;
811 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
812 (*end)++;
813
814 return 0;
815 }
816
817 /*
818 * This function splits the words in buffer->text, stores the list with
819 * newline separator into out, and saves the offsets of the original words
820 * in buffer->orig.
821 */
822 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
823 regex_t *word_regex)
824 {
825 int i, j;
826 long alloc = 0;
827
828 out->size = 0;
829 out->ptr = NULL;
830
831 /* fake an empty "0th" word */
832 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
833 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
834 buffer->orig_nr = 1;
835
836 for (i = 0; i < buffer->text.size; i++) {
837 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
838 return;
839
840 /* store original boundaries */
841 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
842 buffer->orig_alloc);
843 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
844 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
845 buffer->orig_nr++;
846
847 /* store one word */
848 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
849 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
850 out->ptr[out->size + j - i] = '\n';
851 out->size += j - i + 1;
852
853 i = j - 1;
854 }
855 }
856
857 /* this executes the word diff on the accumulated buffers */
858 static void diff_words_show(struct diff_words_data *diff_words)
859 {
860 xpparam_t xpp;
861 xdemitconf_t xecfg;
862 mmfile_t minus, plus;
863 struct diff_words_style *style = diff_words->style;
864
865 struct diff_options *opt = diff_words->opt;
866 struct strbuf *msgbuf;
867 char *line_prefix = "";
868
869 assert(opt);
870 if (opt->output_prefix) {
871 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
872 line_prefix = msgbuf->buf;
873 }
874
875 /* special case: only removal */
876 if (!diff_words->plus.text.size) {
877 fputs(line_prefix, diff_words->opt->file);
878 fn_out_diff_words_write_helper(diff_words->opt->file,
879 &style->old, style->newline,
880 diff_words->minus.text.size,
881 diff_words->minus.text.ptr, line_prefix);
882 diff_words->minus.text.size = 0;
883 return;
884 }
885
886 diff_words->current_plus = diff_words->plus.text.ptr;
887 diff_words->last_minus = 0;
888
889 memset(&xpp, 0, sizeof(xpp));
890 memset(&xecfg, 0, sizeof(xecfg));
891 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
892 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
893 xpp.flags = 0;
894 /* as only the hunk header will be parsed, we need a 0-context */
895 xecfg.ctxlen = 0;
896 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
897 &xpp, &xecfg);
898 free(minus.ptr);
899 free(plus.ptr);
900 if (diff_words->current_plus != diff_words->plus.text.ptr +
901 diff_words->plus.text.size) {
902 if (color_words_output_graph_prefix(diff_words))
903 fputs(line_prefix, diff_words->opt->file);
904 fn_out_diff_words_write_helper(diff_words->opt->file,
905 &style->ctx, style->newline,
906 diff_words->plus.text.ptr + diff_words->plus.text.size
907 - diff_words->current_plus, diff_words->current_plus,
908 line_prefix);
909 }
910 diff_words->minus.text.size = diff_words->plus.text.size = 0;
911 }
912
913 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
914 static void diff_words_flush(struct emit_callback *ecbdata)
915 {
916 if (ecbdata->diff_words->minus.text.size ||
917 ecbdata->diff_words->plus.text.size)
918 diff_words_show(ecbdata->diff_words);
919 }
920
921 static void free_diff_words_data(struct emit_callback *ecbdata)
922 {
923 if (ecbdata->diff_words) {
924 diff_words_flush(ecbdata);
925 free (ecbdata->diff_words->minus.text.ptr);
926 free (ecbdata->diff_words->minus.orig);
927 free (ecbdata->diff_words->plus.text.ptr);
928 free (ecbdata->diff_words->plus.orig);
929 if (ecbdata->diff_words->word_regex) {
930 regfree(ecbdata->diff_words->word_regex);
931 free(ecbdata->diff_words->word_regex);
932 }
933 free(ecbdata->diff_words);
934 ecbdata->diff_words = NULL;
935 }
936 }
937
938 const char *diff_get_color(int diff_use_color, enum color_diff ix)
939 {
940 if (diff_use_color)
941 return diff_colors[ix];
942 return "";
943 }
944
945 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
946 {
947 const char *cp;
948 unsigned long allot;
949 size_t l = len;
950
951 if (ecb->truncate)
952 return ecb->truncate(line, len);
953 cp = line;
954 allot = l;
955 while (0 < l) {
956 (void) utf8_width(&cp, &l);
957 if (!cp)
958 break; /* truncated in the middle? */
959 }
960 return allot - l;
961 }
962
963 static void find_lno(const char *line, struct emit_callback *ecbdata)
964 {
965 const char *p;
966 ecbdata->lno_in_preimage = 0;
967 ecbdata->lno_in_postimage = 0;
968 p = strchr(line, '-');
969 if (!p)
970 return; /* cannot happen */
971 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
972 p = strchr(p, '+');
973 if (!p)
974 return; /* cannot happen */
975 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
976 }
977
978 static void fn_out_consume(void *priv, char *line, unsigned long len)
979 {
980 struct emit_callback *ecbdata = priv;
981 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
982 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
983 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
984 struct diff_options *o = ecbdata->opt;
985 char *line_prefix = "";
986 struct strbuf *msgbuf;
987
988 if (o && o->output_prefix) {
989 msgbuf = o->output_prefix(o, o->output_prefix_data);
990 line_prefix = msgbuf->buf;
991 }
992
993 if (ecbdata->header) {
994 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
995 strbuf_reset(ecbdata->header);
996 ecbdata->header = NULL;
997 }
998 *(ecbdata->found_changesp) = 1;
999
1000 if (ecbdata->label_path[0]) {
1001 const char *name_a_tab, *name_b_tab;
1002
1003 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1004 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1005
1006 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1007 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1008 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1009 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1010 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1011 }
1012
1013 if (diff_suppress_blank_empty
1014 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1015 line[0] = '\n';
1016 len = 1;
1017 }
1018
1019 if (line[0] == '@') {
1020 if (ecbdata->diff_words)
1021 diff_words_flush(ecbdata);
1022 len = sane_truncate_line(ecbdata, line, len);
1023 find_lno(line, ecbdata);
1024 emit_hunk_header(ecbdata, line, len);
1025 if (line[len-1] != '\n')
1026 putc('\n', ecbdata->opt->file);
1027 return;
1028 }
1029
1030 if (len < 1) {
1031 emit_line(ecbdata->opt, reset, reset, line, len);
1032 if (ecbdata->diff_words
1033 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1034 fputs("~\n", ecbdata->opt->file);
1035 return;
1036 }
1037
1038 if (ecbdata->diff_words) {
1039 if (line[0] == '-') {
1040 diff_words_append(line, len,
1041 &ecbdata->diff_words->minus);
1042 return;
1043 } else if (line[0] == '+') {
1044 diff_words_append(line, len,
1045 &ecbdata->diff_words->plus);
1046 return;
1047 }
1048 diff_words_flush(ecbdata);
1049 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1050 emit_line(ecbdata->opt, plain, reset, line, len);
1051 fputs("~\n", ecbdata->opt->file);
1052 } else {
1053 /*
1054 * Skip the prefix character, if any. With
1055 * diff_suppress_blank_empty, there may be
1056 * none.
1057 */
1058 if (line[0] != '\n') {
1059 line++;
1060 len--;
1061 }
1062 emit_line(ecbdata->opt, plain, reset, line, len);
1063 }
1064 return;
1065 }
1066
1067 if (line[0] != '+') {
1068 const char *color =
1069 diff_get_color(ecbdata->color_diff,
1070 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1071 ecbdata->lno_in_preimage++;
1072 if (line[0] == ' ')
1073 ecbdata->lno_in_postimage++;
1074 emit_line(ecbdata->opt, color, reset, line, len);
1075 } else {
1076 ecbdata->lno_in_postimage++;
1077 emit_add_line(reset, ecbdata, line + 1, len - 1);
1078 }
1079 }
1080
1081 static char *pprint_rename(const char *a, const char *b)
1082 {
1083 const char *old = a;
1084 const char *new = b;
1085 struct strbuf name = STRBUF_INIT;
1086 int pfx_length, sfx_length;
1087 int len_a = strlen(a);
1088 int len_b = strlen(b);
1089 int a_midlen, b_midlen;
1090 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1091 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1092
1093 if (qlen_a || qlen_b) {
1094 quote_c_style(a, &name, NULL, 0);
1095 strbuf_addstr(&name, " => ");
1096 quote_c_style(b, &name, NULL, 0);
1097 return strbuf_detach(&name, NULL);
1098 }
1099
1100 /* Find common prefix */
1101 pfx_length = 0;
1102 while (*old && *new && *old == *new) {
1103 if (*old == '/')
1104 pfx_length = old - a + 1;
1105 old++;
1106 new++;
1107 }
1108
1109 /* Find common suffix */
1110 old = a + len_a;
1111 new = b + len_b;
1112 sfx_length = 0;
1113 while (a <= old && b <= new && *old == *new) {
1114 if (*old == '/')
1115 sfx_length = len_a - (old - a);
1116 old--;
1117 new--;
1118 }
1119
1120 /*
1121 * pfx{mid-a => mid-b}sfx
1122 * {pfx-a => pfx-b}sfx
1123 * pfx{sfx-a => sfx-b}
1124 * name-a => name-b
1125 */
1126 a_midlen = len_a - pfx_length - sfx_length;
1127 b_midlen = len_b - pfx_length - sfx_length;
1128 if (a_midlen < 0)
1129 a_midlen = 0;
1130 if (b_midlen < 0)
1131 b_midlen = 0;
1132
1133 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1134 if (pfx_length + sfx_length) {
1135 strbuf_add(&name, a, pfx_length);
1136 strbuf_addch(&name, '{');
1137 }
1138 strbuf_add(&name, a + pfx_length, a_midlen);
1139 strbuf_addstr(&name, " => ");
1140 strbuf_add(&name, b + pfx_length, b_midlen);
1141 if (pfx_length + sfx_length) {
1142 strbuf_addch(&name, '}');
1143 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1144 }
1145 return strbuf_detach(&name, NULL);
1146 }
1147
1148 struct diffstat_t {
1149 int nr;
1150 int alloc;
1151 struct diffstat_file {
1152 char *from_name;
1153 char *name;
1154 char *print_name;
1155 unsigned is_unmerged:1;
1156 unsigned is_binary:1;
1157 unsigned is_renamed:1;
1158 uintmax_t added, deleted;
1159 } **files;
1160 };
1161
1162 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1163 const char *name_a,
1164 const char *name_b)
1165 {
1166 struct diffstat_file *x;
1167 x = xcalloc(sizeof (*x), 1);
1168 if (diffstat->nr == diffstat->alloc) {
1169 diffstat->alloc = alloc_nr(diffstat->alloc);
1170 diffstat->files = xrealloc(diffstat->files,
1171 diffstat->alloc * sizeof(x));
1172 }
1173 diffstat->files[diffstat->nr++] = x;
1174 if (name_b) {
1175 x->from_name = xstrdup(name_a);
1176 x->name = xstrdup(name_b);
1177 x->is_renamed = 1;
1178 }
1179 else {
1180 x->from_name = NULL;
1181 x->name = xstrdup(name_a);
1182 }
1183 return x;
1184 }
1185
1186 static void diffstat_consume(void *priv, char *line, unsigned long len)
1187 {
1188 struct diffstat_t *diffstat = priv;
1189 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1190
1191 if (line[0] == '+')
1192 x->added++;
1193 else if (line[0] == '-')
1194 x->deleted++;
1195 }
1196
1197 const char mime_boundary_leader[] = "------------";
1198
1199 static int scale_linear(int it, int width, int max_change)
1200 {
1201 /*
1202 * make sure that at least one '-' is printed if there were deletions,
1203 * and likewise for '+'.
1204 */
1205 if (max_change < 2)
1206 return it;
1207 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1208 }
1209
1210 static void show_name(FILE *file,
1211 const char *prefix, const char *name, int len)
1212 {
1213 fprintf(file, " %s%-*s |", prefix, len, name);
1214 }
1215
1216 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1217 {
1218 if (cnt <= 0)
1219 return;
1220 fprintf(file, "%s", set);
1221 while (cnt--)
1222 putc(ch, file);
1223 fprintf(file, "%s", reset);
1224 }
1225
1226 static void fill_print_name(struct diffstat_file *file)
1227 {
1228 char *pname;
1229
1230 if (file->print_name)
1231 return;
1232
1233 if (!file->is_renamed) {
1234 struct strbuf buf = STRBUF_INIT;
1235 if (quote_c_style(file->name, &buf, NULL, 0)) {
1236 pname = strbuf_detach(&buf, NULL);
1237 } else {
1238 pname = file->name;
1239 strbuf_release(&buf);
1240 }
1241 } else {
1242 pname = pprint_rename(file->from_name, file->name);
1243 }
1244 file->print_name = pname;
1245 }
1246
1247 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1248 {
1249 int i, len, add, del, adds = 0, dels = 0;
1250 uintmax_t max_change = 0, max_len = 0;
1251 int total_files = data->nr;
1252 int width, name_width;
1253 const char *reset, *add_c, *del_c;
1254 const char *line_prefix = "";
1255 struct strbuf *msg = NULL;
1256
1257 if (data->nr == 0)
1258 return;
1259
1260 if (options->output_prefix) {
1261 msg = options->output_prefix(options, options->output_prefix_data);
1262 line_prefix = msg->buf;
1263 }
1264
1265 width = options->stat_width ? options->stat_width : 80;
1266 name_width = options->stat_name_width ? options->stat_name_width : 50;
1267
1268 /* Sanity: give at least 5 columns to the graph,
1269 * but leave at least 10 columns for the name.
1270 */
1271 if (width < 25)
1272 width = 25;
1273 if (name_width < 10)
1274 name_width = 10;
1275 else if (width < name_width + 15)
1276 name_width = width - 15;
1277
1278 /* Find the longest filename and max number of changes */
1279 reset = diff_get_color_opt(options, DIFF_RESET);
1280 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1281 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1282
1283 for (i = 0; i < data->nr; i++) {
1284 struct diffstat_file *file = data->files[i];
1285 uintmax_t change = file->added + file->deleted;
1286 fill_print_name(file);
1287 len = strlen(file->print_name);
1288 if (max_len < len)
1289 max_len = len;
1290
1291 if (file->is_binary || file->is_unmerged)
1292 continue;
1293 if (max_change < change)
1294 max_change = change;
1295 }
1296
1297 /* Compute the width of the graph part;
1298 * 10 is for one blank at the beginning of the line plus
1299 * " | count " between the name and the graph.
1300 *
1301 * From here on, name_width is the width of the name area,
1302 * and width is the width of the graph area.
1303 */
1304 name_width = (name_width < max_len) ? name_width : max_len;
1305 if (width < (name_width + 10) + max_change)
1306 width = width - (name_width + 10);
1307 else
1308 width = max_change;
1309
1310 for (i = 0; i < data->nr; i++) {
1311 const char *prefix = "";
1312 char *name = data->files[i]->print_name;
1313 uintmax_t added = data->files[i]->added;
1314 uintmax_t deleted = data->files[i]->deleted;
1315 int name_len;
1316
1317 /*
1318 * "scale" the filename
1319 */
1320 len = name_width;
1321 name_len = strlen(name);
1322 if (name_width < name_len) {
1323 char *slash;
1324 prefix = "...";
1325 len -= 3;
1326 name += name_len - len;
1327 slash = strchr(name, '/');
1328 if (slash)
1329 name = slash;
1330 }
1331
1332 if (data->files[i]->is_binary) {
1333 fprintf(options->file, "%s", line_prefix);
1334 show_name(options->file, prefix, name, len);
1335 fprintf(options->file, " Bin ");
1336 fprintf(options->file, "%s%"PRIuMAX"%s",
1337 del_c, deleted, reset);
1338 fprintf(options->file, " -> ");
1339 fprintf(options->file, "%s%"PRIuMAX"%s",
1340 add_c, added, reset);
1341 fprintf(options->file, " bytes");
1342 fprintf(options->file, "\n");
1343 continue;
1344 }
1345 else if (data->files[i]->is_unmerged) {
1346 fprintf(options->file, "%s", line_prefix);
1347 show_name(options->file, prefix, name, len);
1348 fprintf(options->file, " Unmerged\n");
1349 continue;
1350 }
1351 else if (!data->files[i]->is_renamed &&
1352 (added + deleted == 0)) {
1353 total_files--;
1354 continue;
1355 }
1356
1357 /*
1358 * scale the add/delete
1359 */
1360 add = added;
1361 del = deleted;
1362 adds += add;
1363 dels += del;
1364
1365 if (width <= max_change) {
1366 add = scale_linear(add, width, max_change);
1367 del = scale_linear(del, width, max_change);
1368 }
1369 fprintf(options->file, "%s", line_prefix);
1370 show_name(options->file, prefix, name, len);
1371 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1372 added + deleted ? " " : "");
1373 show_graph(options->file, '+', add, add_c, reset);
1374 show_graph(options->file, '-', del, del_c, reset);
1375 fprintf(options->file, "\n");
1376 }
1377 fprintf(options->file, "%s", line_prefix);
1378 fprintf(options->file,
1379 " %d files changed, %d insertions(+), %d deletions(-)\n",
1380 total_files, adds, dels);
1381 }
1382
1383 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1384 {
1385 int i, adds = 0, dels = 0, total_files = data->nr;
1386
1387 if (data->nr == 0)
1388 return;
1389
1390 for (i = 0; i < data->nr; i++) {
1391 if (!data->files[i]->is_binary &&
1392 !data->files[i]->is_unmerged) {
1393 int added = data->files[i]->added;
1394 int deleted= data->files[i]->deleted;
1395 if (!data->files[i]->is_renamed &&
1396 (added + deleted == 0)) {
1397 total_files--;
1398 } else {
1399 adds += added;
1400 dels += deleted;
1401 }
1402 }
1403 }
1404 if (options->output_prefix) {
1405 struct strbuf *msg = NULL;
1406 msg = options->output_prefix(options,
1407 options->output_prefix_data);
1408 fprintf(options->file, "%s", msg->buf);
1409 }
1410 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1411 total_files, adds, dels);
1412 }
1413
1414 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1415 {
1416 int i;
1417
1418 if (data->nr == 0)
1419 return;
1420
1421 for (i = 0; i < data->nr; i++) {
1422 struct diffstat_file *file = data->files[i];
1423
1424 if (options->output_prefix) {
1425 struct strbuf *msg = NULL;
1426 msg = options->output_prefix(options,
1427 options->output_prefix_data);
1428 fprintf(options->file, "%s", msg->buf);
1429 }
1430
1431 if (file->is_binary)
1432 fprintf(options->file, "-\t-\t");
1433 else
1434 fprintf(options->file,
1435 "%"PRIuMAX"\t%"PRIuMAX"\t",
1436 file->added, file->deleted);
1437 if (options->line_termination) {
1438 fill_print_name(file);
1439 if (!file->is_renamed)
1440 write_name_quoted(file->name, options->file,
1441 options->line_termination);
1442 else {
1443 fputs(file->print_name, options->file);
1444 putc(options->line_termination, options->file);
1445 }
1446 } else {
1447 if (file->is_renamed) {
1448 putc('\0', options->file);
1449 write_name_quoted(file->from_name, options->file, '\0');
1450 }
1451 write_name_quoted(file->name, options->file, '\0');
1452 }
1453 }
1454 }
1455
1456 struct dirstat_file {
1457 const char *name;
1458 unsigned long changed;
1459 };
1460
1461 struct dirstat_dir {
1462 struct dirstat_file *files;
1463 int alloc, nr, percent, cumulative;
1464 };
1465
1466 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1467 unsigned long changed, const char *base, int baselen)
1468 {
1469 unsigned long this_dir = 0;
1470 unsigned int sources = 0;
1471 const char *line_prefix = "";
1472 struct strbuf *msg = NULL;
1473
1474 if (opt->output_prefix) {
1475 msg = opt->output_prefix(opt, opt->output_prefix_data);
1476 line_prefix = msg->buf;
1477 }
1478
1479 while (dir->nr) {
1480 struct dirstat_file *f = dir->files;
1481 int namelen = strlen(f->name);
1482 unsigned long this;
1483 char *slash;
1484
1485 if (namelen < baselen)
1486 break;
1487 if (memcmp(f->name, base, baselen))
1488 break;
1489 slash = strchr(f->name + baselen, '/');
1490 if (slash) {
1491 int newbaselen = slash + 1 - f->name;
1492 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1493 sources++;
1494 } else {
1495 this = f->changed;
1496 dir->files++;
1497 dir->nr--;
1498 sources += 2;
1499 }
1500 this_dir += this;
1501 }
1502
1503 /*
1504 * We don't report dirstat's for
1505 * - the top level
1506 * - or cases where everything came from a single directory
1507 * under this directory (sources == 1).
1508 */
1509 if (baselen && sources != 1) {
1510 int permille = this_dir * 1000 / changed;
1511 if (permille) {
1512 int percent = permille / 10;
1513 if (percent >= dir->percent) {
1514 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1515 percent, permille % 10, baselen, base);
1516 if (!dir->cumulative)
1517 return 0;
1518 }
1519 }
1520 }
1521 return this_dir;
1522 }
1523
1524 static int dirstat_compare(const void *_a, const void *_b)
1525 {
1526 const struct dirstat_file *a = _a;
1527 const struct dirstat_file *b = _b;
1528 return strcmp(a->name, b->name);
1529 }
1530
1531 static void show_dirstat(struct diff_options *options)
1532 {
1533 int i;
1534 unsigned long changed;
1535 struct dirstat_dir dir;
1536 struct diff_queue_struct *q = &diff_queued_diff;
1537
1538 dir.files = NULL;
1539 dir.alloc = 0;
1540 dir.nr = 0;
1541 dir.percent = options->dirstat_percent;
1542 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1543
1544 changed = 0;
1545 for (i = 0; i < q->nr; i++) {
1546 struct diff_filepair *p = q->queue[i];
1547 const char *name;
1548 unsigned long copied, added, damage;
1549 int content_changed;
1550
1551 name = p->two->path ? p->two->path : p->one->path;
1552
1553 if (p->one->sha1_valid && p->two->sha1_valid)
1554 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1555 else
1556 content_changed = 1;
1557
1558 if (!content_changed) {
1559 /*
1560 * The SHA1 has not changed, so pre-/post-content is
1561 * identical. We can therefore skip looking at the
1562 * file contents altogether.
1563 */
1564 damage = 0;
1565 goto found_damage;
1566 }
1567
1568 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1569 /*
1570 * In --dirstat-by-file mode, we don't really need to
1571 * look at the actual file contents at all.
1572 * The fact that the SHA1 changed is enough for us to
1573 * add this file to the list of results
1574 * (with each file contributing equal damage).
1575 */
1576 damage = 1;
1577 goto found_damage;
1578 }
1579
1580 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1581 diff_populate_filespec(p->one, 0);
1582 diff_populate_filespec(p->two, 0);
1583 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1584 &copied, &added);
1585 diff_free_filespec_data(p->one);
1586 diff_free_filespec_data(p->two);
1587 } else if (DIFF_FILE_VALID(p->one)) {
1588 diff_populate_filespec(p->one, 1);
1589 copied = added = 0;
1590 diff_free_filespec_data(p->one);
1591 } else if (DIFF_FILE_VALID(p->two)) {
1592 diff_populate_filespec(p->two, 1);
1593 copied = 0;
1594 added = p->two->size;
1595 diff_free_filespec_data(p->two);
1596 } else
1597 continue;
1598
1599 /*
1600 * Original minus copied is the removed material,
1601 * added is the new material. They are both damages
1602 * made to the preimage.
1603 * If the resulting damage is zero, we know that
1604 * diffcore_count_changes() considers the two entries to
1605 * be identical, but since content_changed is true, we
1606 * know that there must have been _some_ kind of change,
1607 * so we force all entries to have damage > 0.
1608 */
1609 damage = (p->one->size - copied) + added;
1610 if (!damage)
1611 damage = 1;
1612
1613 found_damage:
1614 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1615 dir.files[dir.nr].name = name;
1616 dir.files[dir.nr].changed = damage;
1617 changed += damage;
1618 dir.nr++;
1619 }
1620
1621 /* This can happen even with many files, if everything was renames */
1622 if (!changed)
1623 return;
1624
1625 /* Show all directories with more than x% of the changes */
1626 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1627 gather_dirstat(options, &dir, changed, "", 0);
1628 }
1629
1630 static void free_diffstat_info(struct diffstat_t *diffstat)
1631 {
1632 int i;
1633 for (i = 0; i < diffstat->nr; i++) {
1634 struct diffstat_file *f = diffstat->files[i];
1635 if (f->name != f->print_name)
1636 free(f->print_name);
1637 free(f->name);
1638 free(f->from_name);
1639 free(f);
1640 }
1641 free(diffstat->files);
1642 }
1643
1644 struct checkdiff_t {
1645 const char *filename;
1646 int lineno;
1647 int conflict_marker_size;
1648 struct diff_options *o;
1649 unsigned ws_rule;
1650 unsigned status;
1651 };
1652
1653 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1654 {
1655 char firstchar;
1656 int cnt;
1657
1658 if (len < marker_size + 1)
1659 return 0;
1660 firstchar = line[0];
1661 switch (firstchar) {
1662 case '=': case '>': case '<': case '|':
1663 break;
1664 default:
1665 return 0;
1666 }
1667 for (cnt = 1; cnt < marker_size; cnt++)
1668 if (line[cnt] != firstchar)
1669 return 0;
1670 /* line[1] thru line[marker_size-1] are same as firstchar */
1671 if (len < marker_size + 1 || !isspace(line[marker_size]))
1672 return 0;
1673 return 1;
1674 }
1675
1676 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1677 {
1678 struct checkdiff_t *data = priv;
1679 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1680 int marker_size = data->conflict_marker_size;
1681 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1682 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1683 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1684 char *err;
1685 char *line_prefix = "";
1686 struct strbuf *msgbuf;
1687
1688 assert(data->o);
1689 if (data->o->output_prefix) {
1690 msgbuf = data->o->output_prefix(data->o,
1691 data->o->output_prefix_data);
1692 line_prefix = msgbuf->buf;
1693 }
1694
1695 if (line[0] == '+') {
1696 unsigned bad;
1697 data->lineno++;
1698 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1699 data->status |= 1;
1700 fprintf(data->o->file,
1701 "%s%s:%d: leftover conflict marker\n",
1702 line_prefix, data->filename, data->lineno);
1703 }
1704 bad = ws_check(line + 1, len - 1, data->ws_rule);
1705 if (!bad)
1706 return;
1707 data->status |= bad;
1708 err = whitespace_error_string(bad);
1709 fprintf(data->o->file, "%s%s:%d: %s.\n",
1710 line_prefix, data->filename, data->lineno, err);
1711 free(err);
1712 emit_line(data->o, set, reset, line, 1);
1713 ws_check_emit(line + 1, len - 1, data->ws_rule,
1714 data->o->file, set, reset, ws);
1715 } else if (line[0] == ' ') {
1716 data->lineno++;
1717 } else if (line[0] == '@') {
1718 char *plus = strchr(line, '+');
1719 if (plus)
1720 data->lineno = strtol(plus, NULL, 10) - 1;
1721 else
1722 die("invalid diff");
1723 }
1724 }
1725
1726 static unsigned char *deflate_it(char *data,
1727 unsigned long size,
1728 unsigned long *result_size)
1729 {
1730 int bound;
1731 unsigned char *deflated;
1732 z_stream stream;
1733
1734 memset(&stream, 0, sizeof(stream));
1735 deflateInit(&stream, zlib_compression_level);
1736 bound = deflateBound(&stream, size);
1737 deflated = xmalloc(bound);
1738 stream.next_out = deflated;
1739 stream.avail_out = bound;
1740
1741 stream.next_in = (unsigned char *)data;
1742 stream.avail_in = size;
1743 while (deflate(&stream, Z_FINISH) == Z_OK)
1744 ; /* nothing */
1745 deflateEnd(&stream);
1746 *result_size = stream.total_out;
1747 return deflated;
1748 }
1749
1750 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1751 {
1752 void *cp;
1753 void *delta;
1754 void *deflated;
1755 void *data;
1756 unsigned long orig_size;
1757 unsigned long delta_size;
1758 unsigned long deflate_size;
1759 unsigned long data_size;
1760
1761 /* We could do deflated delta, or we could do just deflated two,
1762 * whichever is smaller.
1763 */
1764 delta = NULL;
1765 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1766 if (one->size && two->size) {
1767 delta = diff_delta(one->ptr, one->size,
1768 two->ptr, two->size,
1769 &delta_size, deflate_size);
1770 if (delta) {
1771 void *to_free = delta;
1772 orig_size = delta_size;
1773 delta = deflate_it(delta, delta_size, &delta_size);
1774 free(to_free);
1775 }
1776 }
1777
1778 if (delta && delta_size < deflate_size) {
1779 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1780 free(deflated);
1781 data = delta;
1782 data_size = delta_size;
1783 }
1784 else {
1785 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1786 free(delta);
1787 data = deflated;
1788 data_size = deflate_size;
1789 }
1790
1791 /* emit data encoded in base85 */
1792 cp = data;
1793 while (data_size) {
1794 int bytes = (52 < data_size) ? 52 : data_size;
1795 char line[70];
1796 data_size -= bytes;
1797 if (bytes <= 26)
1798 line[0] = bytes + 'A' - 1;
1799 else
1800 line[0] = bytes - 26 + 'a' - 1;
1801 encode_85(line + 1, cp, bytes);
1802 cp = (char *) cp + bytes;
1803 fprintf(file, "%s", prefix);
1804 fputs(line, file);
1805 fputc('\n', file);
1806 }
1807 fprintf(file, "%s\n", prefix);
1808 free(data);
1809 }
1810
1811 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1812 {
1813 fprintf(file, "%sGIT binary patch\n", prefix);
1814 emit_binary_diff_body(file, one, two, prefix);
1815 emit_binary_diff_body(file, two, one, prefix);
1816 }
1817
1818 static void diff_filespec_load_driver(struct diff_filespec *one)
1819 {
1820 /* Use already-loaded driver */
1821 if (one->driver)
1822 return;
1823
1824 if (S_ISREG(one->mode))
1825 one->driver = userdiff_find_by_path(one->path);
1826
1827 /* Fallback to default settings */
1828 if (!one->driver)
1829 one->driver = userdiff_find_by_name("default");
1830 }
1831
1832 int diff_filespec_is_binary(struct diff_filespec *one)
1833 {
1834 if (one->is_binary == -1) {
1835 diff_filespec_load_driver(one);
1836 if (one->driver->binary != -1)
1837 one->is_binary = one->driver->binary;
1838 else {
1839 if (!one->data && DIFF_FILE_VALID(one))
1840 diff_populate_filespec(one, 0);
1841 if (one->data)
1842 one->is_binary = buffer_is_binary(one->data,
1843 one->size);
1844 if (one->is_binary == -1)
1845 one->is_binary = 0;
1846 }
1847 }
1848 return one->is_binary;
1849 }
1850
1851 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1852 {
1853 diff_filespec_load_driver(one);
1854 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1855 }
1856
1857 static const char *userdiff_word_regex(struct diff_filespec *one)
1858 {
1859 diff_filespec_load_driver(one);
1860 return one->driver->word_regex;
1861 }
1862
1863 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1864 {
1865 if (!options->a_prefix)
1866 options->a_prefix = a;
1867 if (!options->b_prefix)
1868 options->b_prefix = b;
1869 }
1870
1871 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1872 {
1873 if (!DIFF_FILE_VALID(one))
1874 return NULL;
1875
1876 diff_filespec_load_driver(one);
1877 if (!one->driver->textconv)
1878 return NULL;
1879
1880 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1881 struct notes_cache *c = xmalloc(sizeof(*c));
1882 struct strbuf name = STRBUF_INIT;
1883
1884 strbuf_addf(&name, "textconv/%s", one->driver->name);
1885 notes_cache_init(c, name.buf, one->driver->textconv);
1886 one->driver->textconv_cache = c;
1887 }
1888
1889 return one->driver;
1890 }
1891
1892 static void builtin_diff(const char *name_a,
1893 const char *name_b,
1894 struct diff_filespec *one,
1895 struct diff_filespec *two,
1896 const char *xfrm_msg,
1897 int must_show_header,
1898 struct diff_options *o,
1899 int complete_rewrite)
1900 {
1901 mmfile_t mf1, mf2;
1902 const char *lbl[2];
1903 char *a_one, *b_two;
1904 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1905 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1906 const char *a_prefix, *b_prefix;
1907 struct userdiff_driver *textconv_one = NULL;
1908 struct userdiff_driver *textconv_two = NULL;
1909 struct strbuf header = STRBUF_INIT;
1910 struct strbuf *msgbuf;
1911 char *line_prefix = "";
1912
1913 if (o->output_prefix) {
1914 msgbuf = o->output_prefix(o, o->output_prefix_data);
1915 line_prefix = msgbuf->buf;
1916 }
1917
1918 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1919 (!one->mode || S_ISGITLINK(one->mode)) &&
1920 (!two->mode || S_ISGITLINK(two->mode))) {
1921 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1922 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1923 show_submodule_summary(o->file, one ? one->path : two->path,
1924 one->sha1, two->sha1, two->dirty_submodule,
1925 del, add, reset);
1926 return;
1927 }
1928
1929 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1930 textconv_one = get_textconv(one);
1931 textconv_two = get_textconv(two);
1932 }
1933
1934 diff_set_mnemonic_prefix(o, "a/", "b/");
1935 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1936 a_prefix = o->b_prefix;
1937 b_prefix = o->a_prefix;
1938 } else {
1939 a_prefix = o->a_prefix;
1940 b_prefix = o->b_prefix;
1941 }
1942
1943 /* Never use a non-valid filename anywhere if at all possible */
1944 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1945 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1946
1947 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1948 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1949 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1950 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1951 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1952 if (lbl[0][0] == '/') {
1953 /* /dev/null */
1954 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1955 if (xfrm_msg)
1956 strbuf_addstr(&header, xfrm_msg);
1957 must_show_header = 1;
1958 }
1959 else if (lbl[1][0] == '/') {
1960 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1961 if (xfrm_msg)
1962 strbuf_addstr(&header, xfrm_msg);
1963 must_show_header = 1;
1964 }
1965 else {
1966 if (one->mode != two->mode) {
1967 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1968 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1969 must_show_header = 1;
1970 }
1971 if (xfrm_msg)
1972 strbuf_addstr(&header, xfrm_msg);
1973
1974 /*
1975 * we do not run diff between different kind
1976 * of objects.
1977 */
1978 if ((one->mode ^ two->mode) & S_IFMT)
1979 goto free_ab_and_return;
1980 if (complete_rewrite &&
1981 (textconv_one || !diff_filespec_is_binary(one)) &&
1982 (textconv_two || !diff_filespec_is_binary(two))) {
1983 fprintf(o->file, "%s", header.buf);
1984 strbuf_reset(&header);
1985 emit_rewrite_diff(name_a, name_b, one, two,
1986 textconv_one, textconv_two, o);
1987 o->found_changes = 1;
1988 goto free_ab_and_return;
1989 }
1990 }
1991
1992 if (!DIFF_OPT_TST(o, TEXT) &&
1993 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1994 (!textconv_two && diff_filespec_is_binary(two)) )) {
1995 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1996 die("unable to read files to diff");
1997 /* Quite common confusing case */
1998 if (mf1.size == mf2.size &&
1999 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2000 if (must_show_header)
2001 fprintf(o->file, "%s", header.buf);
2002 goto free_ab_and_return;
2003 }
2004 fprintf(o->file, "%s", header.buf);
2005 strbuf_reset(&header);
2006 if (DIFF_OPT_TST(o, BINARY))
2007 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2008 else
2009 fprintf(o->file, "%sBinary files %s and %s differ\n",
2010 line_prefix, lbl[0], lbl[1]);
2011 o->found_changes = 1;
2012 }
2013 else {
2014 /* Crazy xdl interfaces.. */
2015 const char *diffopts = getenv("GIT_DIFF_OPTS");
2016 xpparam_t xpp;
2017 xdemitconf_t xecfg;
2018 struct emit_callback ecbdata;
2019 const struct userdiff_funcname *pe;
2020
2021 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2022 fprintf(o->file, "%s", header.buf);
2023 strbuf_reset(&header);
2024 }
2025
2026 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2027 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2028
2029 pe = diff_funcname_pattern(one);
2030 if (!pe)
2031 pe = diff_funcname_pattern(two);
2032
2033 memset(&xpp, 0, sizeof(xpp));
2034 memset(&xecfg, 0, sizeof(xecfg));
2035 memset(&ecbdata, 0, sizeof(ecbdata));
2036 ecbdata.label_path = lbl;
2037 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2038 ecbdata.found_changesp = &o->found_changes;
2039 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2040 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2041 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2042 ecbdata.opt = o;
2043 ecbdata.header = header.len ? &header : NULL;
2044 xpp.flags = o->xdl_opts;
2045 xecfg.ctxlen = o->context;
2046 xecfg.interhunkctxlen = o->interhunkcontext;
2047 xecfg.flags = XDL_EMIT_FUNCNAMES;
2048 if (pe)
2049 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2050 if (!diffopts)
2051 ;
2052 else if (!prefixcmp(diffopts, "--unified="))
2053 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2054 else if (!prefixcmp(diffopts, "-u"))
2055 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2056 if (o->word_diff) {
2057 int i;
2058
2059 ecbdata.diff_words =
2060 xcalloc(1, sizeof(struct diff_words_data));
2061 ecbdata.diff_words->type = o->word_diff;
2062 ecbdata.diff_words->opt = o;
2063 if (!o->word_regex)
2064 o->word_regex = userdiff_word_regex(one);
2065 if (!o->word_regex)
2066 o->word_regex = userdiff_word_regex(two);
2067 if (!o->word_regex)
2068 o->word_regex = diff_word_regex_cfg;
2069 if (o->word_regex) {
2070 ecbdata.diff_words->word_regex = (regex_t *)
2071 xmalloc(sizeof(regex_t));
2072 if (regcomp(ecbdata.diff_words->word_regex,
2073 o->word_regex,
2074 REG_EXTENDED | REG_NEWLINE))
2075 die ("Invalid regular expression: %s",
2076 o->word_regex);
2077 }
2078 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2079 if (o->word_diff == diff_words_styles[i].type) {
2080 ecbdata.diff_words->style =
2081 &diff_words_styles[i];
2082 break;
2083 }
2084 }
2085 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2086 struct diff_words_style *st = ecbdata.diff_words->style;
2087 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2088 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2089 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2090 }
2091 }
2092 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2093 &xpp, &xecfg);
2094 if (o->word_diff)
2095 free_diff_words_data(&ecbdata);
2096 if (textconv_one)
2097 free(mf1.ptr);
2098 if (textconv_two)
2099 free(mf2.ptr);
2100 xdiff_clear_find_func(&xecfg);
2101 }
2102
2103 free_ab_and_return:
2104 strbuf_release(&header);
2105 diff_free_filespec_data(one);
2106 diff_free_filespec_data(two);
2107 free(a_one);
2108 free(b_two);
2109 return;
2110 }
2111
2112 static void builtin_diffstat(const char *name_a, const char *name_b,
2113 struct diff_filespec *one,
2114 struct diff_filespec *two,
2115 struct diffstat_t *diffstat,
2116 struct diff_options *o,
2117 int complete_rewrite)
2118 {
2119 mmfile_t mf1, mf2;
2120 struct diffstat_file *data;
2121
2122 data = diffstat_add(diffstat, name_a, name_b);
2123
2124 if (!one || !two) {
2125 data->is_unmerged = 1;
2126 return;
2127 }
2128
2129 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2130 data->is_binary = 1;
2131 data->added = diff_filespec_size(two);
2132 data->deleted = diff_filespec_size(one);
2133 }
2134
2135 else if (complete_rewrite) {
2136 diff_populate_filespec(one, 0);
2137 diff_populate_filespec(two, 0);
2138 data->deleted = count_lines(one->data, one->size);
2139 data->added = count_lines(two->data, two->size);
2140 }
2141
2142 else {
2143 /* Crazy xdl interfaces.. */
2144 xpparam_t xpp;
2145 xdemitconf_t xecfg;
2146
2147 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2148 die("unable to read files to diff");
2149
2150 memset(&xpp, 0, sizeof(xpp));
2151 memset(&xecfg, 0, sizeof(xecfg));
2152 xpp.flags = o->xdl_opts;
2153 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2154 &xpp, &xecfg);
2155 }
2156
2157 diff_free_filespec_data(one);
2158 diff_free_filespec_data(two);
2159 }
2160
2161 static void builtin_checkdiff(const char *name_a, const char *name_b,
2162 const char *attr_path,
2163 struct diff_filespec *one,
2164 struct diff_filespec *two,
2165 struct diff_options *o)
2166 {
2167 mmfile_t mf1, mf2;
2168 struct checkdiff_t data;
2169
2170 if (!two)
2171 return;
2172
2173 memset(&data, 0, sizeof(data));
2174 data.filename = name_b ? name_b : name_a;
2175 data.lineno = 0;
2176 data.o = o;
2177 data.ws_rule = whitespace_rule(attr_path);
2178 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2179
2180 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2181 die("unable to read files to diff");
2182
2183 /*
2184 * All the other codepaths check both sides, but not checking
2185 * the "old" side here is deliberate. We are checking the newly
2186 * introduced changes, and as long as the "new" side is text, we
2187 * can and should check what it introduces.
2188 */
2189 if (diff_filespec_is_binary(two))
2190 goto free_and_return;
2191 else {
2192 /* Crazy xdl interfaces.. */
2193 xpparam_t xpp;
2194 xdemitconf_t xecfg;
2195
2196 memset(&xpp, 0, sizeof(xpp));
2197 memset(&xecfg, 0, sizeof(xecfg));
2198 xecfg.ctxlen = 1; /* at least one context line */
2199 xpp.flags = 0;
2200 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2201 &xpp, &xecfg);
2202
2203 if (data.ws_rule & WS_BLANK_AT_EOF) {
2204 struct emit_callback ecbdata;
2205 int blank_at_eof;
2206
2207 ecbdata.ws_rule = data.ws_rule;
2208 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2209 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2210
2211 if (blank_at_eof) {
2212 static char *err;
2213 if (!err)
2214 err = whitespace_error_string(WS_BLANK_AT_EOF);
2215 fprintf(o->file, "%s:%d: %s.\n",
2216 data.filename, blank_at_eof, err);
2217 data.status = 1; /* report errors */
2218 }
2219 }
2220 }
2221 free_and_return:
2222 diff_free_filespec_data(one);
2223 diff_free_filespec_data(two);
2224 if (data.status)
2225 DIFF_OPT_SET(o, CHECK_FAILED);
2226 }
2227
2228 struct diff_filespec *alloc_filespec(const char *path)
2229 {
2230 int namelen = strlen(path);
2231 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2232
2233 memset(spec, 0, sizeof(*spec));
2234 spec->path = (char *)(spec + 1);
2235 memcpy(spec->path, path, namelen+1);
2236 spec->count = 1;
2237 spec->is_binary = -1;
2238 return spec;
2239 }
2240
2241 void free_filespec(struct diff_filespec *spec)
2242 {
2243 if (!--spec->count) {
2244 diff_free_filespec_data(spec);
2245 free(spec);
2246 }
2247 }
2248
2249 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2250 unsigned short mode)
2251 {
2252 if (mode) {
2253 spec->mode = canon_mode(mode);
2254 hashcpy(spec->sha1, sha1);
2255 spec->sha1_valid = !is_null_sha1(sha1);
2256 }
2257 }
2258
2259 /*
2260 * Given a name and sha1 pair, if the index tells us the file in
2261 * the work tree has that object contents, return true, so that
2262 * prepare_temp_file() does not have to inflate and extract.
2263 */
2264 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2265 {
2266 struct cache_entry *ce;
2267 struct stat st;
2268 int pos, len;
2269
2270 /*
2271 * We do not read the cache ourselves here, because the
2272 * benchmark with my previous version that always reads cache
2273 * shows that it makes things worse for diff-tree comparing
2274 * two linux-2.6 kernel trees in an already checked out work
2275 * tree. This is because most diff-tree comparisons deal with
2276 * only a small number of files, while reading the cache is
2277 * expensive for a large project, and its cost outweighs the
2278 * savings we get by not inflating the object to a temporary
2279 * file. Practically, this code only helps when we are used
2280 * by diff-cache --cached, which does read the cache before
2281 * calling us.
2282 */
2283 if (!active_cache)
2284 return 0;
2285
2286 /* We want to avoid the working directory if our caller
2287 * doesn't need the data in a normal file, this system
2288 * is rather slow with its stat/open/mmap/close syscalls,
2289 * and the object is contained in a pack file. The pack
2290 * is probably already open and will be faster to obtain
2291 * the data through than the working directory. Loose
2292 * objects however would tend to be slower as they need
2293 * to be individually opened and inflated.
2294 */
2295 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2296 return 0;
2297
2298 len = strlen(name);
2299 pos = cache_name_pos(name, len);
2300 if (pos < 0)
2301 return 0;
2302 ce = active_cache[pos];
2303
2304 /*
2305 * This is not the sha1 we are looking for, or
2306 * unreusable because it is not a regular file.
2307 */
2308 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2309 return 0;
2310
2311 /*
2312 * If ce is marked as "assume unchanged", there is no
2313 * guarantee that work tree matches what we are looking for.
2314 */
2315 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2316 return 0;
2317
2318 /*
2319 * If ce matches the file in the work tree, we can reuse it.
2320 */
2321 if (ce_uptodate(ce) ||
2322 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2323 return 1;
2324
2325 return 0;
2326 }
2327
2328 static int populate_from_stdin(struct diff_filespec *s)
2329 {
2330 struct strbuf buf = STRBUF_INIT;
2331 size_t size = 0;
2332
2333 if (strbuf_read(&buf, 0, 0) < 0)
2334 return error("error while reading from stdin %s",
2335 strerror(errno));
2336
2337 s->should_munmap = 0;
2338 s->data = strbuf_detach(&buf, &size);
2339 s->size = size;
2340 s->should_free = 1;
2341 return 0;
2342 }
2343
2344 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2345 {
2346 int len;
2347 char *data = xmalloc(100), *dirty = "";
2348
2349 /* Are we looking at the work tree? */
2350 if (s->dirty_submodule)
2351 dirty = "-dirty";
2352
2353 len = snprintf(data, 100,
2354 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2355 s->data = data;
2356 s->size = len;
2357 s->should_free = 1;
2358 if (size_only) {
2359 s->data = NULL;
2360 free(data);
2361 }
2362 return 0;
2363 }
2364
2365 /*
2366 * While doing rename detection and pickaxe operation, we may need to
2367 * grab the data for the blob (or file) for our own in-core comparison.
2368 * diff_filespec has data and size fields for this purpose.
2369 */
2370 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2371 {
2372 int err = 0;
2373 if (!DIFF_FILE_VALID(s))
2374 die("internal error: asking to populate invalid file.");
2375 if (S_ISDIR(s->mode))
2376 return -1;
2377
2378 if (s->data)
2379 return 0;
2380
2381 if (size_only && 0 < s->size)
2382 return 0;
2383
2384 if (S_ISGITLINK(s->mode))
2385 return diff_populate_gitlink(s, size_only);
2386
2387 if (!s->sha1_valid ||
2388 reuse_worktree_file(s->path, s->sha1, 0)) {
2389 struct strbuf buf = STRBUF_INIT;
2390 struct stat st;
2391 int fd;
2392
2393 if (!strcmp(s->path, "-"))
2394 return populate_from_stdin(s);
2395
2396 if (lstat(s->path, &st) < 0) {
2397 if (errno == ENOENT) {
2398 err_empty:
2399 err = -1;
2400 empty:
2401 s->data = (char *)"";
2402 s->size = 0;
2403 return err;
2404 }
2405 }
2406 s->size = xsize_t(st.st_size);
2407 if (!s->size)
2408 goto empty;
2409 if (S_ISLNK(st.st_mode)) {
2410 struct strbuf sb = STRBUF_INIT;
2411
2412 if (strbuf_readlink(&sb, s->path, s->size))
2413 goto err_empty;
2414 s->size = sb.len;
2415 s->data = strbuf_detach(&sb, NULL);
2416 s->should_free = 1;
2417 return 0;
2418 }
2419 if (size_only)
2420 return 0;
2421 fd = open(s->path, O_RDONLY);
2422 if (fd < 0)
2423 goto err_empty;
2424 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2425 close(fd);
2426 s->should_munmap = 1;
2427
2428 /*
2429 * Convert from working tree format to canonical git format
2430 */
2431 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2432 size_t size = 0;
2433 munmap(s->data, s->size);
2434 s->should_munmap = 0;
2435 s->data = strbuf_detach(&buf, &size);
2436 s->size = size;
2437 s->should_free = 1;
2438 }
2439 }
2440 else {
2441 enum object_type type;
2442 if (size_only) {
2443 type = sha1_object_info(s->sha1, &s->size);
2444 if (type < 0)
2445 die("unable to read %s", sha1_to_hex(s->sha1));
2446 } else {
2447 s->data = read_sha1_file(s->sha1, &type, &s->size);
2448 if (!s->data)
2449 die("unable to read %s", sha1_to_hex(s->sha1));
2450 s->should_free = 1;
2451 }
2452 }
2453 return 0;
2454 }
2455
2456 void diff_free_filespec_blob(struct diff_filespec *s)
2457 {
2458 if (s->should_free)
2459 free(s->data);
2460 else if (s->should_munmap)
2461 munmap(s->data, s->size);
2462
2463 if (s->should_free || s->should_munmap) {
2464 s->should_free = s->should_munmap = 0;
2465 s->data = NULL;
2466 }
2467 }
2468
2469 void diff_free_filespec_data(struct diff_filespec *s)
2470 {
2471 diff_free_filespec_blob(s);
2472 free(s->cnt_data);
2473 s->cnt_data = NULL;
2474 }
2475
2476 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2477 void *blob,
2478 unsigned long size,
2479 const unsigned char *sha1,
2480 int mode)
2481 {
2482 int fd;
2483 struct strbuf buf = STRBUF_INIT;
2484 struct strbuf template = STRBUF_INIT;
2485 char *path_dup = xstrdup(path);
2486 const char *base = basename(path_dup);
2487
2488 /* Generate "XXXXXX_basename.ext" */
2489 strbuf_addstr(&template, "XXXXXX_");
2490 strbuf_addstr(&template, base);
2491
2492 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2493 strlen(base) + 1);
2494 if (fd < 0)
2495 die_errno("unable to create temp-file");
2496 if (convert_to_working_tree(path,
2497 (const char *)blob, (size_t)size, &buf)) {
2498 blob = buf.buf;
2499 size = buf.len;
2500 }
2501 if (write_in_full(fd, blob, size) != size)
2502 die_errno("unable to write temp-file");
2503 close(fd);
2504 temp->name = temp->tmp_path;
2505 strcpy(temp->hex, sha1_to_hex(sha1));
2506 temp->hex[40] = 0;
2507 sprintf(temp->mode, "%06o", mode);
2508 strbuf_release(&buf);
2509 strbuf_release(&template);
2510 free(path_dup);
2511 }
2512
2513 static struct diff_tempfile *prepare_temp_file(const char *name,
2514 struct diff_filespec *one)
2515 {
2516 struct diff_tempfile *temp = claim_diff_tempfile();
2517
2518 if (!DIFF_FILE_VALID(one)) {
2519 not_a_valid_file:
2520 /* A '-' entry produces this for file-2, and
2521 * a '+' entry produces this for file-1.
2522 */
2523 temp->name = "/dev/null";
2524 strcpy(temp->hex, ".");
2525 strcpy(temp->mode, ".");
2526 return temp;
2527 }
2528
2529 if (!remove_tempfile_installed) {
2530 atexit(remove_tempfile);
2531 sigchain_push_common(remove_tempfile_on_signal);
2532 remove_tempfile_installed = 1;
2533 }
2534
2535 if (!one->sha1_valid ||
2536 reuse_worktree_file(name, one->sha1, 1)) {
2537 struct stat st;
2538 if (lstat(name, &st) < 0) {
2539 if (errno == ENOENT)
2540 goto not_a_valid_file;
2541 die_errno("stat(%s)", name);
2542 }
2543 if (S_ISLNK(st.st_mode)) {
2544 struct strbuf sb = STRBUF_INIT;
2545 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2546 die_errno("readlink(%s)", name);
2547 prep_temp_blob(name, temp, sb.buf, sb.len,
2548 (one->sha1_valid ?
2549 one->sha1 : null_sha1),
2550 (one->sha1_valid ?
2551 one->mode : S_IFLNK));
2552 strbuf_release(&sb);
2553 }
2554 else {
2555 /* we can borrow from the file in the work tree */
2556 temp->name = name;
2557 if (!one->sha1_valid)
2558 strcpy(temp->hex, sha1_to_hex(null_sha1));
2559 else
2560 strcpy(temp->hex, sha1_to_hex(one->sha1));
2561 /* Even though we may sometimes borrow the
2562 * contents from the work tree, we always want
2563 * one->mode. mode is trustworthy even when
2564 * !(one->sha1_valid), as long as
2565 * DIFF_FILE_VALID(one).
2566 */
2567 sprintf(temp->mode, "%06o", one->mode);
2568 }
2569 return temp;
2570 }
2571 else {
2572 if (diff_populate_filespec(one, 0))
2573 die("cannot read data blob for %s", one->path);
2574 prep_temp_blob(name, temp, one->data, one->size,
2575 one->sha1, one->mode);
2576 }
2577 return temp;
2578 }
2579
2580 /* An external diff command takes:
2581 *
2582 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2583 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2584 *
2585 */
2586 static void run_external_diff(const char *pgm,
2587 const char *name,
2588 const char *other,
2589 struct diff_filespec *one,
2590 struct diff_filespec *two,
2591 const char *xfrm_msg,
2592 int complete_rewrite)
2593 {
2594 const char *spawn_arg[10];
2595 int retval;
2596 const char **arg = &spawn_arg[0];
2597
2598 if (one && two) {
2599 struct diff_tempfile *temp_one, *temp_two;
2600 const char *othername = (other ? other : name);
2601 temp_one = prepare_temp_file(name, one);
2602 temp_two = prepare_temp_file(othername, two);
2603 *arg++ = pgm;
2604 *arg++ = name;
2605 *arg++ = temp_one->name;
2606 *arg++ = temp_one->hex;
2607 *arg++ = temp_one->mode;
2608 *arg++ = temp_two->name;
2609 *arg++ = temp_two->hex;
2610 *arg++ = temp_two->mode;
2611 if (other) {
2612 *arg++ = other;
2613 *arg++ = xfrm_msg;
2614 }
2615 } else {
2616 *arg++ = pgm;
2617 *arg++ = name;
2618 }
2619 *arg = NULL;
2620 fflush(NULL);
2621 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2622 remove_tempfile();
2623 if (retval) {
2624 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2625 exit(1);
2626 }
2627 }
2628
2629 static int similarity_index(struct diff_filepair *p)
2630 {
2631 return p->score * 100 / MAX_SCORE;
2632 }
2633
2634 static void fill_metainfo(struct strbuf *msg,
2635 const char *name,
2636 const char *other,
2637 struct diff_filespec *one,
2638 struct diff_filespec *two,
2639 struct diff_options *o,
2640 struct diff_filepair *p,
2641 int *must_show_header,
2642 int use_color)
2643 {
2644 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2645 const char *reset = diff_get_color(use_color, DIFF_RESET);
2646 struct strbuf *msgbuf;
2647 char *line_prefix = "";
2648
2649 *must_show_header = 1;
2650 if (o->output_prefix) {
2651 msgbuf = o->output_prefix(o, o->output_prefix_data);
2652 line_prefix = msgbuf->buf;
2653 }
2654 strbuf_init(msg, PATH_MAX * 2 + 300);
2655 switch (p->status) {
2656 case DIFF_STATUS_COPIED:
2657 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2658 line_prefix, set, similarity_index(p));
2659 strbuf_addf(msg, "%s\n%s%scopy from ",
2660 reset, line_prefix, set);
2661 quote_c_style(name, msg, NULL, 0);
2662 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2663 quote_c_style(other, msg, NULL, 0);
2664 strbuf_addf(msg, "%s\n", reset);
2665 break;
2666 case DIFF_STATUS_RENAMED:
2667 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2668 line_prefix, set, similarity_index(p));
2669 strbuf_addf(msg, "%s\n%s%srename from ",
2670 reset, line_prefix, set);
2671 quote_c_style(name, msg, NULL, 0);
2672 strbuf_addf(msg, "%s\n%s%srename to ",
2673 reset, line_prefix, set);
2674 quote_c_style(other, msg, NULL, 0);
2675 strbuf_addf(msg, "%s\n", reset);
2676 break;
2677 case DIFF_STATUS_MODIFIED:
2678 if (p->score) {
2679 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2680 line_prefix,
2681 set, similarity_index(p), reset);
2682 break;
2683 }
2684 /* fallthru */
2685 default:
2686 *must_show_header = 0;
2687 }
2688 if (one && two && hashcmp(one->sha1, two->sha1)) {
2689 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2690
2691 if (DIFF_OPT_TST(o, BINARY)) {
2692 mmfile_t mf;
2693 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2694 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2695 abbrev = 40;
2696 }
2697 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2698 find_unique_abbrev(one->sha1, abbrev));
2699 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2700 if (one->mode == two->mode)
2701 strbuf_addf(msg, " %06o", one->mode);
2702 strbuf_addf(msg, "%s\n", reset);
2703 }
2704 }
2705
2706 static void run_diff_cmd(const char *pgm,
2707 const char *name,
2708 const char *other,
2709 const char *attr_path,
2710 struct diff_filespec *one,
2711 struct diff_filespec *two,
2712 struct strbuf *msg,
2713 struct diff_options *o,
2714 struct diff_filepair *p)
2715 {
2716 const char *xfrm_msg = NULL;
2717 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2718 int must_show_header = 0;
2719
2720 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2721 pgm = NULL;
2722 else {
2723 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2724 if (drv && drv->external)
2725 pgm = drv->external;
2726 }
2727
2728 if (msg) {
2729 /*
2730 * don't use colors when the header is intended for an
2731 * external diff driver
2732 */
2733 fill_metainfo(msg, name, other, one, two, o, p,
2734 &must_show_header,
2735 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2736 xfrm_msg = msg->len ? msg->buf : NULL;
2737 }
2738
2739 if (pgm) {
2740 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2741 complete_rewrite);
2742 return;
2743 }
2744 if (one && two)
2745 builtin_diff(name, other ? other : name,
2746 one, two, xfrm_msg, must_show_header,
2747 o, complete_rewrite);
2748 else
2749 fprintf(o->file, "* Unmerged path %s\n", name);
2750 }
2751
2752 static void diff_fill_sha1_info(struct diff_filespec *one)
2753 {
2754 if (DIFF_FILE_VALID(one)) {
2755 if (!one->sha1_valid) {
2756 struct stat st;
2757 if (!strcmp(one->path, "-")) {
2758 hashcpy(one->sha1, null_sha1);
2759 return;
2760 }
2761 if (lstat(one->path, &st) < 0)
2762 die_errno("stat '%s'", one->path);
2763 if (index_path(one->sha1, one->path, &st, 0))
2764 die("cannot hash %s", one->path);
2765 }
2766 }
2767 else
2768 hashclr(one->sha1);
2769 }
2770
2771 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2772 {
2773 /* Strip the prefix but do not molest /dev/null and absolute paths */
2774 if (*namep && **namep != '/') {
2775 *namep += prefix_length;
2776 if (**namep == '/')
2777 ++*namep;
2778 }
2779 if (*otherp && **otherp != '/') {
2780 *otherp += prefix_length;
2781 if (**otherp == '/')
2782 ++*otherp;
2783 }
2784 }
2785
2786 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2787 {
2788 const char *pgm = external_diff();
2789 struct strbuf msg;
2790 struct diff_filespec *one = p->one;
2791 struct diff_filespec *two = p->two;
2792 const char *name;
2793 const char *other;
2794 const char *attr_path;
2795
2796 name = p->one->path;
2797 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2798 attr_path = name;
2799 if (o->prefix_length)
2800 strip_prefix(o->prefix_length, &name, &other);
2801
2802 if (DIFF_PAIR_UNMERGED(p)) {
2803 run_diff_cmd(pgm, name, NULL, attr_path,
2804 NULL, NULL, NULL, o, p);
2805 return;
2806 }
2807
2808 diff_fill_sha1_info(one);
2809 diff_fill_sha1_info(two);
2810
2811 if (!pgm &&
2812 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2813 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2814 /*
2815 * a filepair that changes between file and symlink
2816 * needs to be split into deletion and creation.
2817 */
2818 struct diff_filespec *null = alloc_filespec(two->path);
2819 run_diff_cmd(NULL, name, other, attr_path,
2820 one, null, &msg, o, p);
2821 free(null);
2822 strbuf_release(&msg);
2823
2824 null = alloc_filespec(one->path);
2825 run_diff_cmd(NULL, name, other, attr_path,
2826 null, two, &msg, o, p);
2827 free(null);
2828 }
2829 else
2830 run_diff_cmd(pgm, name, other, attr_path,
2831 one, two, &msg, o, p);
2832
2833 strbuf_release(&msg);
2834 }
2835
2836 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2837 struct diffstat_t *diffstat)
2838 {
2839 const char *name;
2840 const char *other;
2841 int complete_rewrite = 0;
2842
2843 if (DIFF_PAIR_UNMERGED(p)) {
2844 /* unmerged */
2845 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2846 return;
2847 }
2848
2849 name = p->one->path;
2850 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2851
2852 if (o->prefix_length)
2853 strip_prefix(o->prefix_length, &name, &other);
2854
2855 diff_fill_sha1_info(p->one);
2856 diff_fill_sha1_info(p->two);
2857
2858 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2859 complete_rewrite = 1;
2860 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2861 }
2862
2863 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2864 {
2865 const char *name;
2866 const char *other;
2867 const char *attr_path;
2868
2869 if (DIFF_PAIR_UNMERGED(p)) {
2870 /* unmerged */
2871 return;
2872 }
2873
2874 name = p->one->path;
2875 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2876 attr_path = other ? other : name;
2877
2878 if (o->prefix_length)
2879 strip_prefix(o->prefix_length, &name, &other);
2880
2881 diff_fill_sha1_info(p->one);
2882 diff_fill_sha1_info(p->two);
2883
2884 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2885 }
2886
2887 void diff_setup(struct diff_options *options)
2888 {
2889 memcpy(options, &default_diff_options, sizeof(*options));
2890
2891 options->file = stdout;
2892
2893 options->line_termination = '\n';
2894 options->break_opt = -1;
2895 options->rename_limit = -1;
2896 options->dirstat_percent = 3;
2897 options->context = 3;
2898
2899 options->change = diff_change;
2900 options->add_remove = diff_addremove;
2901 if (diff_use_color_default > 0)
2902 DIFF_OPT_SET(options, COLOR_DIFF);
2903 options->detect_rename = diff_detect_rename_default;
2904
2905 if (diff_no_prefix) {
2906 options->a_prefix = options->b_prefix = "";
2907 } else if (!diff_mnemonic_prefix) {
2908 options->a_prefix = "a/";
2909 options->b_prefix = "b/";
2910 }
2911 }
2912
2913 int diff_setup_done(struct diff_options *options)
2914 {
2915 int count = 0;
2916
2917 if (options->output_format & DIFF_FORMAT_NAME)
2918 count++;
2919 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2920 count++;
2921 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2922 count++;
2923 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2924 count++;
2925 if (count > 1)
2926 die("--name-only, --name-status, --check and -s are mutually exclusive");
2927
2928 /*
2929 * Most of the time we can say "there are changes"
2930 * only by checking if there are changed paths, but
2931 * --ignore-whitespace* options force us to look
2932 * inside contents.
2933 */
2934
2935 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2936 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2937 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2938 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2939 else
2940 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2941
2942 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2943 options->detect_rename = DIFF_DETECT_COPY;
2944
2945 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2946 options->prefix = NULL;
2947 if (options->prefix)
2948 options->prefix_length = strlen(options->prefix);
2949 else
2950 options->prefix_length = 0;
2951
2952 if (options->output_format & (DIFF_FORMAT_NAME |
2953 DIFF_FORMAT_NAME_STATUS |
2954 DIFF_FORMAT_CHECKDIFF |
2955 DIFF_FORMAT_NO_OUTPUT))
2956 options->output_format &= ~(DIFF_FORMAT_RAW |
2957 DIFF_FORMAT_NUMSTAT |
2958 DIFF_FORMAT_DIFFSTAT |
2959 DIFF_FORMAT_SHORTSTAT |
2960 DIFF_FORMAT_DIRSTAT |
2961 DIFF_FORMAT_SUMMARY |
2962 DIFF_FORMAT_PATCH);
2963
2964 /*
2965 * These cases always need recursive; we do not drop caller-supplied
2966 * recursive bits for other formats here.
2967 */
2968 if (options->output_format & (DIFF_FORMAT_PATCH |
2969 DIFF_FORMAT_NUMSTAT |
2970 DIFF_FORMAT_DIFFSTAT |
2971 DIFF_FORMAT_SHORTSTAT |
2972 DIFF_FORMAT_DIRSTAT |
2973 DIFF_FORMAT_SUMMARY |
2974 DIFF_FORMAT_CHECKDIFF))
2975 DIFF_OPT_SET(options, RECURSIVE);
2976 /*
2977 * Also pickaxe would not work very well if you do not say recursive
2978 */
2979 if (options->pickaxe)
2980 DIFF_OPT_SET(options, RECURSIVE);
2981 /*
2982 * When patches are generated, submodules diffed against the work tree
2983 * must be checked for dirtiness too so it can be shown in the output
2984 */
2985 if (options->output_format & DIFF_FORMAT_PATCH)
2986 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2987
2988 if (options->detect_rename && options->rename_limit < 0)
2989 options->rename_limit = diff_rename_limit_default;
2990 if (options->setup & DIFF_SETUP_USE_CACHE) {
2991 if (!active_cache)
2992 /* read-cache does not die even when it fails
2993 * so it is safe for us to do this here. Also
2994 * it does not smudge active_cache or active_nr
2995 * when it fails, so we do not have to worry about
2996 * cleaning it up ourselves either.
2997 */
2998 read_cache();
2999 }
3000 if (options->abbrev <= 0 || 40 < options->abbrev)
3001 options->abbrev = 40; /* full */
3002
3003 /*
3004 * It does not make sense to show the first hit we happened
3005 * to have found. It does not make sense not to return with
3006 * exit code in such a case either.
3007 */
3008 if (DIFF_OPT_TST(options, QUICK)) {
3009 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3010 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3011 }
3012
3013 return 0;
3014 }
3015
3016 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3017 {
3018 char c, *eq;
3019 int len;
3020
3021 if (*arg != '-')
3022 return 0;
3023 c = *++arg;
3024 if (!c)
3025 return 0;
3026 if (c == arg_short) {
3027 c = *++arg;
3028 if (!c)
3029 return 1;
3030 if (val && isdigit(c)) {
3031 char *end;
3032 int n = strtoul(arg, &end, 10);
3033 if (*end)
3034 return 0;
3035 *val = n;
3036 return 1;
3037 }
3038 return 0;
3039 }
3040 if (c != '-')
3041 return 0;
3042 arg++;
3043 eq = strchr(arg, '=');
3044 if (eq)
3045 len = eq - arg;
3046 else
3047 len = strlen(arg);
3048 if (!len || strncmp(arg, arg_long, len))
3049 return 0;
3050 if (eq) {
3051 int n;
3052 char *end;
3053 if (!isdigit(*++eq))
3054 return 0;
3055 n = strtoul(eq, &end, 10);
3056 if (*end)
3057 return 0;
3058 *val = n;
3059 }
3060 return 1;
3061 }
3062
3063 static int diff_scoreopt_parse(const char *opt);
3064
3065 static inline int short_opt(char opt, const char **argv,
3066 const char **optarg)
3067 {
3068 const char *arg = argv[0];
3069 if (arg[0] != '-' || arg[1] != opt)
3070 return 0;
3071 if (arg[2] != '\0') {
3072 *optarg = arg + 2;
3073 return 1;
3074 }
3075 if (!argv[1])
3076 die("Option '%c' requires a value", opt);
3077 *optarg = argv[1];
3078 return 2;
3079 }
3080
3081 int parse_long_opt(const char *opt, const char **argv,
3082 const char **optarg)
3083 {
3084 const char *arg = argv[0];
3085 if (arg[0] != '-' || arg[1] != '-')
3086 return 0;
3087 arg += strlen("--");
3088 if (prefixcmp(arg, opt))
3089 return 0;
3090 arg += strlen(opt);
3091 if (*arg == '=') { /* sticked form: --option=value */
3092 *optarg = arg + 1;
3093 return 1;
3094 }
3095 if (*arg != '\0')
3096 return 0;
3097 /* separate form: --option value */
3098 if (!argv[1])
3099 die("Option '--%s' requires a value", opt);
3100 *optarg = argv[1];
3101 return 2;
3102 }
3103
3104 static int stat_opt(struct diff_options *options, const char **av)
3105 {
3106 const char *arg = av[0];
3107 char *end;
3108 int width = options->stat_width;
3109 int name_width = options->stat_name_width;
3110 int argcount = 1;
3111
3112 arg += strlen("--stat");
3113 end = (char *)arg;
3114
3115 switch (*arg) {
3116 case '-':
3117 if (!prefixcmp(arg, "-width")) {
3118 arg += strlen("-width");
3119 if (*arg == '=')
3120 width = strtoul(arg + 1, &end, 10);
3121 else if (!*arg && !av[1])
3122 die("Option '--stat-width' requires a value");
3123 else if (!*arg) {
3124 width = strtoul(av[1], &end, 10);
3125 argcount = 2;
3126 }
3127 } else if (!prefixcmp(arg, "-name-width")) {
3128 arg += strlen("-name-width");
3129 if (*arg == '=')
3130 name_width = strtoul(arg + 1, &end, 10);
3131 else if (!*arg && !av[1])
3132 die("Option '--stat-name-width' requires a value");
3133 else if (!*arg) {
3134 name_width = strtoul(av[1], &end, 10);
3135 argcount = 2;
3136 }
3137 }
3138 break;
3139 case '=':
3140 width = strtoul(arg+1, &end, 10);
3141 if (*end == ',')
3142 name_width = strtoul(end+1, &end, 10);
3143 }
3144
3145 /* Important! This checks all the error cases! */
3146 if (*end)
3147 return 0;
3148 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3149 options->stat_name_width = name_width;
3150 options->stat_width = width;
3151 return argcount;
3152 }
3153
3154 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3155 {
3156 const char *arg = av[0];
3157 const char *optarg;
3158 int argcount;
3159
3160 /* Output format options */
3161 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3162 options->output_format |= DIFF_FORMAT_PATCH;
3163 else if (opt_arg(arg, 'U', "unified", &options->context))
3164 options->output_format |= DIFF_FORMAT_PATCH;
3165 else if (!strcmp(arg, "--raw"))
3166 options->output_format |= DIFF_FORMAT_RAW;
3167 else if (!strcmp(arg, "--patch-with-raw"))
3168 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3169 else if (!strcmp(arg, "--numstat"))
3170 options->output_format |= DIFF_FORMAT_NUMSTAT;
3171 else if (!strcmp(arg, "--shortstat"))
3172 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3173 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3174 options->output_format |= DIFF_FORMAT_DIRSTAT;
3175 else if (!strcmp(arg, "--cumulative")) {
3176 options->output_format |= DIFF_FORMAT_DIRSTAT;
3177 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3178 } else if (opt_arg(arg, 0, "dirstat-by-file",
3179 &options->dirstat_percent)) {
3180 options->output_format |= DIFF_FORMAT_DIRSTAT;
3181 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3182 }
3183 else if (!strcmp(arg, "--check"))
3184 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3185 else if (!strcmp(arg, "--summary"))
3186 options->output_format |= DIFF_FORMAT_SUMMARY;
3187 else if (!strcmp(arg, "--patch-with-stat"))
3188 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3189 else if (!strcmp(arg, "--name-only"))
3190 options->output_format |= DIFF_FORMAT_NAME;
3191 else if (!strcmp(arg, "--name-status"))
3192 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3193 else if (!strcmp(arg, "-s"))
3194 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3195 else if (!prefixcmp(arg, "--stat"))
3196 /* --stat, --stat-width, or --stat-name-width */
3197 return stat_opt(options, av);
3198
3199 /* renames options */
3200 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3201 !strcmp(arg, "--break-rewrites")) {
3202 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3203 return error("invalid argument to -B: %s", arg+2);
3204 }
3205 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3206 !strcmp(arg, "--find-renames")) {
3207 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3208 return error("invalid argument to -M: %s", arg+2);
3209 options->detect_rename = DIFF_DETECT_RENAME;
3210 }
3211 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3212 !strcmp(arg, "--find-copies")) {
3213 if (options->detect_rename == DIFF_DETECT_COPY)
3214 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3215 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3216 return error("invalid argument to -C: %s", arg+2);
3217 options->detect_rename = DIFF_DETECT_COPY;
3218 }
3219 else if (!strcmp(arg, "--no-renames"))
3220 options->detect_rename = 0;
3221 else if (!strcmp(arg, "--relative"))
3222 DIFF_OPT_SET(options, RELATIVE_NAME);
3223 else if (!prefixcmp(arg, "--relative=")) {
3224 DIFF_OPT_SET(options, RELATIVE_NAME);
3225 options->prefix = arg + 11;
3226 }
3227
3228 /* xdiff options */
3229 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3230 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3231 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3232 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3233 else if (!strcmp(arg, "--ignore-space-at-eol"))
3234 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3235 else if (!strcmp(arg, "--patience"))
3236 DIFF_XDL_SET(options, PATIENCE_DIFF);
3237
3238 /* flags options */
3239 else if (!strcmp(arg, "--binary")) {
3240 options->output_format |= DIFF_FORMAT_PATCH;
3241 DIFF_OPT_SET(options, BINARY);
3242 }
3243 else if (!strcmp(arg, "--full-index"))
3244 DIFF_OPT_SET(options, FULL_INDEX);
3245 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3246 DIFF_OPT_SET(options, TEXT);
3247 else if (!strcmp(arg, "-R"))
3248 DIFF_OPT_SET(options, REVERSE_DIFF);
3249 else if (!strcmp(arg, "--find-copies-harder"))
3250 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3251 else if (!strcmp(arg, "--follow"))
3252 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3253 else if (!strcmp(arg, "--color"))
3254 DIFF_OPT_SET(options, COLOR_DIFF);
3255 else if (!prefixcmp(arg, "--color=")) {
3256 int value = git_config_colorbool(NULL, arg+8, -1);
3257 if (value == 0)
3258 DIFF_OPT_CLR(options, COLOR_DIFF);
3259 else if (value > 0)
3260 DIFF_OPT_SET(options, COLOR_DIFF);
3261 else
3262 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3263 }
3264 else if (!strcmp(arg, "--no-color"))
3265 DIFF_OPT_CLR(options, COLOR_DIFF);
3266 else if (!strcmp(arg, "--color-words")) {
3267 DIFF_OPT_SET(options, COLOR_DIFF);
3268 options->word_diff = DIFF_WORDS_COLOR;
3269 }
3270 else if (!prefixcmp(arg, "--color-words=")) {
3271 DIFF_OPT_SET(options, COLOR_DIFF);
3272 options->word_diff = DIFF_WORDS_COLOR;
3273 options->word_regex = arg + 14;
3274 }
3275 else if (!strcmp(arg, "--word-diff")) {
3276 if (options->word_diff == DIFF_WORDS_NONE)
3277 options->word_diff = DIFF_WORDS_PLAIN;
3278 }
3279 else if (!prefixcmp(arg, "--word-diff=")) {
3280 const char *type = arg + 12;
3281 if (!strcmp(type, "plain"))
3282 options->word_diff = DIFF_WORDS_PLAIN;
3283 else if (!strcmp(type, "color")) {
3284 DIFF_OPT_SET(options, COLOR_DIFF);
3285 options->word_diff = DIFF_WORDS_COLOR;
3286 }
3287 else if (!strcmp(type, "porcelain"))
3288 options->word_diff = DIFF_WORDS_PORCELAIN;
3289 else if (!strcmp(type, "none"))
3290 options->word_diff = DIFF_WORDS_NONE;
3291 else
3292 die("bad --word-diff argument: %s", type);
3293 }
3294 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3295 if (options->word_diff == DIFF_WORDS_NONE)
3296 options->word_diff = DIFF_WORDS_PLAIN;
3297 options->word_regex = optarg;
3298 return argcount;
3299 }
3300 else if (!strcmp(arg, "--exit-code"))
3301 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3302 else if (!strcmp(arg, "--quiet"))
3303 DIFF_OPT_SET(options, QUICK);
3304 else if (!strcmp(arg, "--ext-diff"))
3305 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3306 else if (!strcmp(arg, "--no-ext-diff"))
3307 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3308 else if (!strcmp(arg, "--textconv"))
3309 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3310 else if (!strcmp(arg, "--no-textconv"))
3311 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3312 else if (!strcmp(arg, "--ignore-submodules")) {
3313 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3314 handle_ignore_submodules_arg(options, "all");
3315 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3316 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3317 handle_ignore_submodules_arg(options, arg + 20);
3318 } else if (!strcmp(arg, "--submodule"))
3319 DIFF_OPT_SET(options, SUBMODULE_LOG);
3320 else if (!prefixcmp(arg, "--submodule=")) {
3321 if (!strcmp(arg + 12, "log"))
3322 DIFF_OPT_SET(options, SUBMODULE_LOG);
3323 }
3324
3325 /* misc options */
3326 else if (!strcmp(arg, "-z"))
3327 options->line_termination = 0;
3328 else if ((argcount = short_opt('l', av, &optarg))) {
3329 options->rename_limit = strtoul(optarg, NULL, 10);
3330 return argcount;
3331 }
3332 else if ((argcount = short_opt('S', av, &optarg))) {
3333 options->pickaxe = optarg;
3334 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3335 return argcount;
3336 } else if ((argcount = short_opt('G', av, &optarg))) {
3337 options->pickaxe = optarg;
3338 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3339 return argcount;
3340 }
3341 else if (!strcmp(arg, "--pickaxe-all"))
3342 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3343 else if (!strcmp(arg, "--pickaxe-regex"))
3344 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3345 else if ((argcount = short_opt('O', av, &optarg))) {
3346 options->orderfile = optarg;
3347 return argcount;
3348 }
3349 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3350 options->filter = optarg;
3351 return argcount;
3352 }
3353 else if (!strcmp(arg, "--abbrev"))
3354 options->abbrev = DEFAULT_ABBREV;
3355 else if (!prefixcmp(arg, "--abbrev=")) {
3356 options->abbrev = strtoul(arg + 9, NULL, 10);
3357 if (options->abbrev < MINIMUM_ABBREV)
3358 options->abbrev = MINIMUM_ABBREV;
3359 else if (40 < options->abbrev)
3360 options->abbrev = 40;
3361 }
3362 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3363 options->a_prefix = optarg;
3364 return argcount;
3365 }
3366 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3367 options->b_prefix = optarg;
3368 return argcount;
3369 }
3370 else if (!strcmp(arg, "--no-prefix"))
3371 options->a_prefix = options->b_prefix = "";
3372 else if (opt_arg(arg, '\0', "inter-hunk-context",
3373 &options->interhunkcontext))
3374 ;
3375 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3376 options->file = fopen(optarg, "w");
3377 if (!options->file)
3378 die_errno("Could not open '%s'", optarg);
3379 options->close_file = 1;
3380 return argcount;
3381 } else
3382 return 0;
3383 return 1;
3384 }
3385
3386 int parse_rename_score(const char **cp_p)
3387 {
3388 unsigned long num, scale;
3389 int ch, dot;
3390 const char *cp = *cp_p;
3391
3392 num = 0;
3393 scale = 1;
3394 dot = 0;
3395 for (;;) {
3396 ch = *cp;
3397 if ( !dot && ch == '.' ) {
3398 scale = 1;
3399 dot = 1;
3400 } else if ( ch == '%' ) {
3401 scale = dot ? scale*100 : 100;
3402 cp++; /* % is always at the end */
3403 break;
3404 } else if ( ch >= '0' && ch <= '9' ) {
3405 if ( scale < 100000 ) {
3406 scale *= 10;
3407 num = (num*10) + (ch-'0');
3408 }
3409 } else {
3410 break;
3411 }
3412 cp++;
3413 }
3414 *cp_p = cp;
3415
3416 /* user says num divided by scale and we say internally that
3417 * is MAX_SCORE * num / scale.
3418 */
3419 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3420 }
3421
3422 static int diff_scoreopt_parse(const char *opt)
3423 {
3424 int opt1, opt2, cmd;
3425
3426 if (*opt++ != '-')
3427 return -1;
3428 cmd = *opt++;
3429 if (cmd == '-') {
3430 /* convert the long-form arguments into short-form versions */
3431 if (!prefixcmp(opt, "break-rewrites")) {
3432 opt += strlen("break-rewrites");
3433 if (*opt == 0 || *opt++ == '=')
3434 cmd = 'B';
3435 } else if (!prefixcmp(opt, "find-copies")) {
3436 opt += strlen("find-copies");
3437 if (*opt == 0 || *opt++ == '=')
3438 cmd = 'C';
3439 } else if (!prefixcmp(opt, "find-renames")) {
3440 opt += strlen("find-renames");
3441 if (*opt == 0 || *opt++ == '=')
3442 cmd = 'M';
3443 }
3444 }
3445 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3446 return -1; /* that is not a -M, -C nor -B option */
3447
3448 opt1 = parse_rename_score(&opt);
3449 if (cmd != 'B')
3450 opt2 = 0;
3451 else {
3452 if (*opt == 0)
3453 opt2 = 0;
3454 else if (*opt != '/')
3455 return -1; /* we expect -B80/99 or -B80 */
3456 else {
3457 opt++;
3458 opt2 = parse_rename_score(&opt);
3459 }
3460 }
3461 if (*opt != 0)
3462 return -1;
3463 return opt1 | (opt2 << 16);
3464 }
3465
3466 struct diff_queue_struct diff_queued_diff;
3467
3468 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3469 {
3470 if (queue->alloc <= queue->nr) {
3471 queue->alloc = alloc_nr(queue->alloc);
3472 queue->queue = xrealloc(queue->queue,
3473 sizeof(dp) * queue->alloc);
3474 }
3475 queue->queue[queue->nr++] = dp;
3476 }
3477
3478 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3479 struct diff_filespec *one,
3480 struct diff_filespec *two)
3481 {
3482 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3483 dp->one = one;
3484 dp->two = two;
3485 if (queue)
3486 diff_q(queue, dp);
3487 return dp;
3488 }
3489
3490 void diff_free_filepair(struct diff_filepair *p)
3491 {
3492 free_filespec(p->one);
3493 free_filespec(p->two);
3494 free(p);
3495 }
3496
3497 /* This is different from find_unique_abbrev() in that
3498 * it stuffs the result with dots for alignment.
3499 */
3500 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3501 {
3502 int abblen;
3503 const char *abbrev;
3504 if (len == 40)
3505 return sha1_to_hex(sha1);
3506
3507 abbrev = find_unique_abbrev(sha1, len);
3508 abblen = strlen(abbrev);
3509 if (abblen < 37) {
3510 static char hex[41];
3511 if (len < abblen && abblen <= len + 2)
3512 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3513 else
3514 sprintf(hex, "%s...", abbrev);
3515 return hex;
3516 }
3517 return sha1_to_hex(sha1);
3518 }
3519
3520 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3521 {
3522 int line_termination = opt->line_termination;
3523 int inter_name_termination = line_termination ? '\t' : '\0';
3524 if (opt->output_prefix) {
3525 struct strbuf *msg = NULL;
3526 msg = opt->output_prefix(opt, opt->output_prefix_data);
3527 fprintf(opt->file, "%s", msg->buf);
3528 }
3529
3530 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3531 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3532 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3533 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3534 }
3535 if (p->score) {
3536 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3537 inter_name_termination);
3538 } else {
3539 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3540 }
3541
3542 if (p->status == DIFF_STATUS_COPIED ||
3543 p->status == DIFF_STATUS_RENAMED) {
3544 const char *name_a, *name_b;
3545 name_a = p->one->path;
3546 name_b = p->two->path;
3547 strip_prefix(opt->prefix_length, &name_a, &name_b);
3548 write_name_quoted(name_a, opt->file, inter_name_termination);
3549 write_name_quoted(name_b, opt->file, line_termination);
3550 } else {
3551 const char *name_a, *name_b;
3552 name_a = p->one->mode ? p->one->path : p->two->path;
3553 name_b = NULL;
3554 strip_prefix(opt->prefix_length, &name_a, &name_b);
3555 write_name_quoted(name_a, opt->file, line_termination);
3556 }
3557 }
3558
3559 int diff_unmodified_pair(struct diff_filepair *p)
3560 {
3561 /* This function is written stricter than necessary to support
3562 * the currently implemented transformers, but the idea is to
3563 * let transformers to produce diff_filepairs any way they want,
3564 * and filter and clean them up here before producing the output.
3565 */
3566 struct diff_filespec *one = p->one, *two = p->two;
3567
3568 if (DIFF_PAIR_UNMERGED(p))
3569 return 0; /* unmerged is interesting */
3570
3571 /* deletion, addition, mode or type change
3572 * and rename are all interesting.
3573 */
3574 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3575 DIFF_PAIR_MODE_CHANGED(p) ||
3576 strcmp(one->path, two->path))
3577 return 0;
3578
3579 /* both are valid and point at the same path. that is, we are
3580 * dealing with a change.
3581 */
3582 if (one->sha1_valid && two->sha1_valid &&
3583 !hashcmp(one->sha1, two->sha1) &&
3584 !one->dirty_submodule && !two->dirty_submodule)
3585 return 1; /* no change */
3586 if (!one->sha1_valid && !two->sha1_valid)
3587 return 1; /* both look at the same file on the filesystem. */
3588 return 0;
3589 }
3590
3591 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3592 {
3593 if (diff_unmodified_pair(p))
3594 return;
3595
3596 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3597 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3598 return; /* no tree diffs in patch format */
3599
3600 run_diff(p, o);
3601 }
3602
3603 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3604 struct diffstat_t *diffstat)
3605 {
3606 if (diff_unmodified_pair(p))
3607 return;
3608
3609 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3610 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3611 return; /* no useful stat for tree diffs */
3612
3613 run_diffstat(p, o, diffstat);
3614 }
3615
3616 static void diff_flush_checkdiff(struct diff_filepair *p,
3617 struct diff_options *o)
3618 {
3619 if (diff_unmodified_pair(p))
3620 return;
3621
3622 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3623 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3624 return; /* nothing to check in tree diffs */
3625
3626 run_checkdiff(p, o);
3627 }
3628
3629 int diff_queue_is_empty(void)
3630 {
3631 struct diff_queue_struct *q = &diff_queued_diff;
3632 int i;
3633 for (i = 0; i < q->nr; i++)
3634 if (!diff_unmodified_pair(q->queue[i]))
3635 return 0;
3636 return 1;
3637 }
3638
3639 #if DIFF_DEBUG
3640 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3641 {
3642 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3643 x, one ? one : "",
3644 s->path,
3645 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3646 s->mode,
3647 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3648 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3649 x, one ? one : "",
3650 s->size, s->xfrm_flags);
3651 }
3652
3653 void diff_debug_filepair(const struct diff_filepair *p, int i)
3654 {
3655 diff_debug_filespec(p->one, i, "one");
3656 diff_debug_filespec(p->two, i, "two");
3657 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3658 p->score, p->status ? p->status : '?',
3659 p->one->rename_used, p->broken_pair);
3660 }
3661
3662 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3663 {
3664 int i;
3665 if (msg)
3666 fprintf(stderr, "%s\n", msg);
3667 fprintf(stderr, "q->nr = %d\n", q->nr);
3668 for (i = 0; i < q->nr; i++) {
3669 struct diff_filepair *p = q->queue[i];
3670 diff_debug_filepair(p, i);
3671 }
3672 }
3673 #endif
3674
3675 static void diff_resolve_rename_copy(void)
3676 {
3677 int i;
3678 struct diff_filepair *p;
3679 struct diff_queue_struct *q = &diff_queued_diff;
3680
3681 diff_debug_queue("resolve-rename-copy", q);
3682
3683 for (i = 0; i < q->nr; i++) {
3684 p = q->queue[i];
3685 p->status = 0; /* undecided */
3686 if (DIFF_PAIR_UNMERGED(p))
3687 p->status = DIFF_STATUS_UNMERGED;
3688 else if (!DIFF_FILE_VALID(p->one))
3689 p->status = DIFF_STATUS_ADDED;
3690 else if (!DIFF_FILE_VALID(p->two))
3691 p->status = DIFF_STATUS_DELETED;
3692 else if (DIFF_PAIR_TYPE_CHANGED(p))
3693 p->status = DIFF_STATUS_TYPE_CHANGED;
3694
3695 /* from this point on, we are dealing with a pair
3696 * whose both sides are valid and of the same type, i.e.
3697 * either in-place edit or rename/copy edit.
3698 */
3699 else if (DIFF_PAIR_RENAME(p)) {
3700 /*
3701 * A rename might have re-connected a broken
3702 * pair up, causing the pathnames to be the
3703 * same again. If so, that's not a rename at
3704 * all, just a modification..
3705 *
3706 * Otherwise, see if this source was used for
3707 * multiple renames, in which case we decrement
3708 * the count, and call it a copy.
3709 */
3710 if (!strcmp(p->one->path, p->two->path))
3711 p->status = DIFF_STATUS_MODIFIED;
3712 else if (--p->one->rename_used > 0)
3713 p->status = DIFF_STATUS_COPIED;
3714 else
3715 p->status = DIFF_STATUS_RENAMED;
3716 }
3717 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3718 p->one->mode != p->two->mode ||
3719 p->one->dirty_submodule ||
3720 p->two->dirty_submodule ||
3721 is_null_sha1(p->one->sha1))
3722 p->status = DIFF_STATUS_MODIFIED;
3723 else {
3724 /* This is a "no-change" entry and should not
3725 * happen anymore, but prepare for broken callers.
3726 */
3727 error("feeding unmodified %s to diffcore",
3728 p->one->path);
3729 p->status = DIFF_STATUS_UNKNOWN;
3730 }
3731 }
3732 diff_debug_queue("resolve-rename-copy done", q);
3733 }
3734
3735 static int check_pair_status(struct diff_filepair *p)
3736 {
3737 switch (p->status) {
3738 case DIFF_STATUS_UNKNOWN:
3739 return 0;
3740 case 0:
3741 die("internal error in diff-resolve-rename-copy");
3742 default:
3743 return 1;
3744 }
3745 }
3746
3747 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3748 {
3749 int fmt = opt->output_format;
3750
3751 if (fmt & DIFF_FORMAT_CHECKDIFF)
3752 diff_flush_checkdiff(p, opt);
3753 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3754 diff_flush_raw(p, opt);
3755 else if (fmt & DIFF_FORMAT_NAME) {
3756 const char *name_a, *name_b;
3757 name_a = p->two->path;
3758 name_b = NULL;
3759 strip_prefix(opt->prefix_length, &name_a, &name_b);
3760 write_name_quoted(name_a, opt->file, opt->line_termination);
3761 }
3762 }
3763
3764 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3765 {
3766 if (fs->mode)
3767 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3768 else
3769 fprintf(file, " %s ", newdelete);
3770 write_name_quoted(fs->path, file, '\n');
3771 }
3772
3773
3774 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3775 const char *line_prefix)
3776 {
3777 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3778 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3779 p->two->mode, show_name ? ' ' : '\n');
3780 if (show_name) {
3781 write_name_quoted(p->two->path, file, '\n');
3782 }
3783 }
3784 }
3785
3786 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3787 const char *line_prefix)
3788 {
3789 char *names = pprint_rename(p->one->path, p->two->path);
3790
3791 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3792 free(names);
3793 show_mode_change(file, p, 0, line_prefix);
3794 }
3795
3796 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3797 {
3798 FILE *file = opt->file;
3799 char *line_prefix = "";
3800
3801 if (opt->output_prefix) {
3802 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3803 line_prefix = buf->buf;
3804 }
3805
3806 switch(p->status) {
3807 case DIFF_STATUS_DELETED:
3808 fputs(line_prefix, file);
3809 show_file_mode_name(file, "delete", p->one);
3810 break;
3811 case DIFF_STATUS_ADDED:
3812 fputs(line_prefix, file);
3813 show_file_mode_name(file, "create", p->two);
3814 break;
3815 case DIFF_STATUS_COPIED:
3816 fputs(line_prefix, file);
3817 show_rename_copy(file, "copy", p, line_prefix);
3818 break;
3819 case DIFF_STATUS_RENAMED:
3820 fputs(line_prefix, file);
3821 show_rename_copy(file, "rename", p, line_prefix);
3822 break;
3823 default:
3824 if (p->score) {
3825 fprintf(file, "%s rewrite ", line_prefix);
3826 write_name_quoted(p->two->path, file, ' ');
3827 fprintf(file, "(%d%%)\n", similarity_index(p));
3828 }
3829 show_mode_change(file, p, !p->score, line_prefix);
3830 break;
3831 }
3832 }
3833
3834 struct patch_id_t {
3835 git_SHA_CTX *ctx;
3836 int patchlen;
3837 };
3838
3839 static int remove_space(char *line, int len)
3840 {
3841 int i;
3842 char *dst = line;
3843 unsigned char c;
3844
3845 for (i = 0; i < len; i++)
3846 if (!isspace((c = line[i])))
3847 *dst++ = c;
3848
3849 return dst - line;
3850 }
3851
3852 static void patch_id_consume(void *priv, char *line, unsigned long len)
3853 {
3854 struct patch_id_t *data = priv;
3855 int new_len;
3856
3857 /* Ignore line numbers when computing the SHA1 of the patch */
3858 if (!prefixcmp(line, "@@ -"))
3859 return;
3860
3861 new_len = remove_space(line, len);
3862
3863 git_SHA1_Update(data->ctx, line, new_len);
3864 data->patchlen += new_len;
3865 }
3866
3867 /* returns 0 upon success, and writes result into sha1 */
3868 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3869 {
3870 struct diff_queue_struct *q = &diff_queued_diff;
3871 int i;
3872 git_SHA_CTX ctx;
3873 struct patch_id_t data;
3874 char buffer[PATH_MAX * 4 + 20];
3875
3876 git_SHA1_Init(&ctx);
3877 memset(&data, 0, sizeof(struct patch_id_t));
3878 data.ctx = &ctx;
3879
3880 for (i = 0; i < q->nr; i++) {
3881 xpparam_t xpp;
3882 xdemitconf_t xecfg;
3883 mmfile_t mf1, mf2;
3884 struct diff_filepair *p = q->queue[i];
3885 int len1, len2;
3886
3887 memset(&xpp, 0, sizeof(xpp));
3888 memset(&xecfg, 0, sizeof(xecfg));
3889 if (p->status == 0)
3890 return error("internal diff status error");
3891 if (p->status == DIFF_STATUS_UNKNOWN)
3892 continue;
3893 if (diff_unmodified_pair(p))
3894 continue;
3895 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3896 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3897 continue;
3898 if (DIFF_PAIR_UNMERGED(p))
3899 continue;
3900
3901 diff_fill_sha1_info(p->one);
3902 diff_fill_sha1_info(p->two);
3903 if (fill_mmfile(&mf1, p->one) < 0 ||
3904 fill_mmfile(&mf2, p->two) < 0)
3905 return error("unable to read files to diff");
3906
3907 len1 = remove_space(p->one->path, strlen(p->one->path));
3908 len2 = remove_space(p->two->path, strlen(p->two->path));
3909 if (p->one->mode == 0)
3910 len1 = snprintf(buffer, sizeof(buffer),
3911 "diff--gita/%.*sb/%.*s"
3912 "newfilemode%06o"
3913 "---/dev/null"
3914 "+++b/%.*s",
3915 len1, p->one->path,
3916 len2, p->two->path,
3917 p->two->mode,
3918 len2, p->two->path);
3919 else if (p->two->mode == 0)
3920 len1 = snprintf(buffer, sizeof(buffer),
3921 "diff--gita/%.*sb/%.*s"
3922 "deletedfilemode%06o"
3923 "---a/%.*s"
3924 "+++/dev/null",
3925 len1, p->one->path,
3926 len2, p->two->path,
3927 p->one->mode,
3928 len1, p->one->path);
3929 else
3930 len1 = snprintf(buffer, sizeof(buffer),
3931 "diff--gita/%.*sb/%.*s"
3932 "---a/%.*s"
3933 "+++b/%.*s",
3934 len1, p->one->path,
3935 len2, p->two->path,
3936 len1, p->one->path,
3937 len2, p->two->path);
3938 git_SHA1_Update(&ctx, buffer, len1);
3939
3940 if (diff_filespec_is_binary(p->one) ||
3941 diff_filespec_is_binary(p->two)) {
3942 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3943 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3944 continue;
3945 }
3946
3947 xpp.flags = 0;
3948 xecfg.ctxlen = 3;
3949 xecfg.flags = 0;
3950 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3951 &xpp, &xecfg);
3952 }
3953
3954 git_SHA1_Final(sha1, &ctx);
3955 return 0;
3956 }
3957
3958 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3959 {
3960 struct diff_queue_struct *q = &diff_queued_diff;
3961 int i;
3962 int result = diff_get_patch_id(options, sha1);
3963
3964 for (i = 0; i < q->nr; i++)
3965 diff_free_filepair(q->queue[i]);
3966
3967 free(q->queue);
3968 DIFF_QUEUE_CLEAR(q);
3969
3970 return result;
3971 }
3972
3973 static int is_summary_empty(const struct diff_queue_struct *q)
3974 {
3975 int i;
3976
3977 for (i = 0; i < q->nr; i++) {
3978 const struct diff_filepair *p = q->queue[i];
3979
3980 switch (p->status) {
3981 case DIFF_STATUS_DELETED:
3982 case DIFF_STATUS_ADDED:
3983 case DIFF_STATUS_COPIED:
3984 case DIFF_STATUS_RENAMED:
3985 return 0;
3986 default:
3987 if (p->score)
3988 return 0;
3989 if (p->one->mode && p->two->mode &&
3990 p->one->mode != p->two->mode)
3991 return 0;
3992 break;
3993 }
3994 }
3995 return 1;
3996 }
3997
3998 static const char rename_limit_warning[] =
3999 "inexact rename detection was skipped due to too many files.";
4000
4001 static const char degrade_cc_to_c_warning[] =
4002 "only found copies from modified paths due to too many files.";
4003
4004 static const char rename_limit_advice[] =
4005 "you may want to set your %s variable to at least "
4006 "%d and retry the command.";
4007
4008 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4009 {
4010 if (degraded_cc)
4011 warning(degrade_cc_to_c_warning);
4012 else if (needed)
4013 warning(rename_limit_warning);
4014 else
4015 return;
4016 if (0 < needed && needed < 32767)
4017 warning(rename_limit_advice, varname, needed);
4018 }
4019
4020 void diff_flush(struct diff_options *options)
4021 {
4022 struct diff_queue_struct *q = &diff_queued_diff;
4023 int i, output_format = options->output_format;
4024 int separator = 0;
4025
4026 /*
4027 * Order: raw, stat, summary, patch
4028 * or: name/name-status/checkdiff (other bits clear)
4029 */
4030 if (!q->nr)
4031 goto free_queue;
4032
4033 if (output_format & (DIFF_FORMAT_RAW |
4034 DIFF_FORMAT_NAME |
4035 DIFF_FORMAT_NAME_STATUS |
4036 DIFF_FORMAT_CHECKDIFF)) {
4037 for (i = 0; i < q->nr; i++) {
4038 struct diff_filepair *p = q->queue[i];
4039 if (check_pair_status(p))
4040 flush_one_pair(p, options);
4041 }
4042 separator++;
4043 }
4044
4045 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4046 struct diffstat_t diffstat;
4047
4048 memset(&diffstat, 0, sizeof(struct diffstat_t));
4049 for (i = 0; i < q->nr; i++) {
4050 struct diff_filepair *p = q->queue[i];
4051 if (check_pair_status(p))
4052 diff_flush_stat(p, options, &diffstat);
4053 }
4054 if (output_format & DIFF_FORMAT_NUMSTAT)
4055 show_numstat(&diffstat, options);
4056 if (output_format & DIFF_FORMAT_DIFFSTAT)
4057 show_stats(&diffstat, options);
4058 if (output_format & DIFF_FORMAT_SHORTSTAT)
4059 show_shortstats(&diffstat, options);
4060 free_diffstat_info(&diffstat);
4061 separator++;
4062 }
4063 if (output_format & DIFF_FORMAT_DIRSTAT)
4064 show_dirstat(options);
4065
4066 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4067 for (i = 0; i < q->nr; i++) {
4068 diff_summary(options, q->queue[i]);
4069 }
4070 separator++;
4071 }
4072
4073 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4074 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4075 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4076 /*
4077 * run diff_flush_patch for the exit status. setting
4078 * options->file to /dev/null should be safe, becaue we
4079 * aren't supposed to produce any output anyway.
4080 */
4081 if (options->close_file)
4082 fclose(options->file);
4083 options->file = fopen("/dev/null", "w");
4084 if (!options->file)
4085 die_errno("Could not open /dev/null");
4086 options->close_file = 1;
4087 for (i = 0; i < q->nr; i++) {
4088 struct diff_filepair *p = q->queue[i];
4089 if (check_pair_status(p))
4090 diff_flush_patch(p, options);
4091 if (options->found_changes)
4092 break;
4093 }
4094 }
4095
4096 if (output_format & DIFF_FORMAT_PATCH) {
4097 if (separator) {
4098 putc(options->line_termination, options->file);
4099 if (options->stat_sep) {
4100 /* attach patch instead of inline */
4101 fputs(options->stat_sep, options->file);
4102 }
4103 }
4104
4105 for (i = 0; i < q->nr; i++) {
4106 struct diff_filepair *p = q->queue[i];
4107 if (check_pair_status(p))
4108 diff_flush_patch(p, options);
4109 }
4110 }
4111
4112 if (output_format & DIFF_FORMAT_CALLBACK)
4113 options->format_callback(q, options, options->format_callback_data);
4114
4115 for (i = 0; i < q->nr; i++)
4116 diff_free_filepair(q->queue[i]);
4117 free_queue:
4118 free(q->queue);
4119 DIFF_QUEUE_CLEAR(q);
4120 if (options->close_file)
4121 fclose(options->file);
4122
4123 /*
4124 * Report the content-level differences with HAS_CHANGES;
4125 * diff_addremove/diff_change does not set the bit when
4126 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4127 */
4128 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4129 if (options->found_changes)
4130 DIFF_OPT_SET(options, HAS_CHANGES);
4131 else
4132 DIFF_OPT_CLR(options, HAS_CHANGES);
4133 }
4134 }
4135
4136 static void diffcore_apply_filter(const char *filter)
4137 {
4138 int i;
4139 struct diff_queue_struct *q = &diff_queued_diff;
4140 struct diff_queue_struct outq;
4141 DIFF_QUEUE_CLEAR(&outq);
4142
4143 if (!filter)
4144 return;
4145
4146 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4147 int found;
4148 for (i = found = 0; !found && i < q->nr; i++) {
4149 struct diff_filepair *p = q->queue[i];
4150 if (((p->status == DIFF_STATUS_MODIFIED) &&
4151 ((p->score &&
4152 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4153 (!p->score &&
4154 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4155 ((p->status != DIFF_STATUS_MODIFIED) &&
4156 strchr(filter, p->status)))
4157 found++;
4158 }
4159 if (found)
4160 return;
4161
4162 /* otherwise we will clear the whole queue
4163 * by copying the empty outq at the end of this
4164 * function, but first clear the current entries
4165 * in the queue.
4166 */
4167 for (i = 0; i < q->nr; i++)
4168 diff_free_filepair(q->queue[i]);
4169 }
4170 else {
4171 /* Only the matching ones */
4172 for (i = 0; i < q->nr; i++) {
4173 struct diff_filepair *p = q->queue[i];
4174
4175 if (((p->status == DIFF_STATUS_MODIFIED) &&
4176 ((p->score &&
4177 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4178 (!p->score &&
4179 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4180 ((p->status != DIFF_STATUS_MODIFIED) &&
4181 strchr(filter, p->status)))
4182 diff_q(&outq, p);
4183 else
4184 diff_free_filepair(p);
4185 }
4186 }
4187 free(q->queue);
4188 *q = outq;
4189 }
4190
4191 /* Check whether two filespecs with the same mode and size are identical */
4192 static int diff_filespec_is_identical(struct diff_filespec *one,
4193 struct diff_filespec *two)
4194 {
4195 if (S_ISGITLINK(one->mode))
4196 return 0;
4197 if (diff_populate_filespec(one, 0))
4198 return 0;
4199 if (diff_populate_filespec(two, 0))
4200 return 0;
4201 return !memcmp(one->data, two->data, one->size);
4202 }
4203
4204 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4205 {
4206 int i;
4207 struct diff_queue_struct *q = &diff_queued_diff;
4208 struct diff_queue_struct outq;
4209 DIFF_QUEUE_CLEAR(&outq);
4210
4211 for (i = 0; i < q->nr; i++) {
4212 struct diff_filepair *p = q->queue[i];
4213
4214 /*
4215 * 1. Entries that come from stat info dirtiness
4216 * always have both sides (iow, not create/delete),
4217 * one side of the object name is unknown, with
4218 * the same mode and size. Keep the ones that
4219 * do not match these criteria. They have real
4220 * differences.
4221 *
4222 * 2. At this point, the file is known to be modified,
4223 * with the same mode and size, and the object
4224 * name of one side is unknown. Need to inspect
4225 * the identical contents.
4226 */
4227 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4228 !DIFF_FILE_VALID(p->two) ||
4229 (p->one->sha1_valid && p->two->sha1_valid) ||
4230 (p->one->mode != p->two->mode) ||
4231 diff_populate_filespec(p->one, 1) ||
4232 diff_populate_filespec(p->two, 1) ||
4233 (p->one->size != p->two->size) ||
4234 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4235 diff_q(&outq, p);
4236 else {
4237 /*
4238 * The caller can subtract 1 from skip_stat_unmatch
4239 * to determine how many paths were dirty only
4240 * due to stat info mismatch.
4241 */
4242 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4243 diffopt->skip_stat_unmatch++;
4244 diff_free_filepair(p);
4245 }
4246 }
4247 free(q->queue);
4248 *q = outq;
4249 }
4250
4251 static int diffnamecmp(const void *a_, const void *b_)
4252 {
4253 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4254 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4255 const char *name_a, *name_b;
4256
4257 name_a = a->one ? a->one->path : a->two->path;
4258 name_b = b->one ? b->one->path : b->two->path;
4259 return strcmp(name_a, name_b);
4260 }
4261
4262 void diffcore_fix_diff_index(struct diff_options *options)
4263 {
4264 struct diff_queue_struct *q = &diff_queued_diff;
4265 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4266 }
4267
4268 void diffcore_std(struct diff_options *options)
4269 {
4270 if (options->skip_stat_unmatch)
4271 diffcore_skip_stat_unmatch(options);
4272 if (!options->found_follow) {
4273 /* See try_to_follow_renames() in tree-diff.c */
4274 if (options->break_opt != -1)
4275 diffcore_break(options->break_opt);
4276 if (options->detect_rename)
4277 diffcore_rename(options);
4278 if (options->break_opt != -1)
4279 diffcore_merge_broken();
4280 }
4281 if (options->pickaxe)
4282 diffcore_pickaxe(options);
4283 if (options->orderfile)
4284 diffcore_order(options->orderfile);
4285 if (!options->found_follow)
4286 /* See try_to_follow_renames() in tree-diff.c */
4287 diff_resolve_rename_copy();
4288 diffcore_apply_filter(options->filter);
4289
4290 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4291 DIFF_OPT_SET(options, HAS_CHANGES);
4292 else
4293 DIFF_OPT_CLR(options, HAS_CHANGES);
4294
4295 options->found_follow = 0;
4296 }
4297
4298 int diff_result_code(struct diff_options *opt, int status)
4299 {
4300 int result = 0;
4301
4302 diff_warn_rename_limit("diff.renamelimit",
4303 opt->needed_rename_limit,
4304 opt->degraded_cc_to_c);
4305 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4306 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4307 return status;
4308 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4309 DIFF_OPT_TST(opt, HAS_CHANGES))
4310 result |= 01;
4311 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4312 DIFF_OPT_TST(opt, CHECK_FAILED))
4313 result |= 02;
4314 return result;
4315 }
4316
4317 /*
4318 * Shall changes to this submodule be ignored?
4319 *
4320 * Submodule changes can be configured to be ignored separately for each path,
4321 * but that configuration can be overridden from the command line.
4322 */
4323 static int is_submodule_ignored(const char *path, struct diff_options *options)
4324 {
4325 int ignored = 0;
4326 unsigned orig_flags = options->flags;
4327 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4328 set_diffopt_flags_from_submodule_config(options, path);
4329 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4330 ignored = 1;
4331 options->flags = orig_flags;
4332 return ignored;
4333 }
4334
4335 void diff_addremove(struct diff_options *options,
4336 int addremove, unsigned mode,
4337 const unsigned char *sha1,
4338 const char *concatpath, unsigned dirty_submodule)
4339 {
4340 struct diff_filespec *one, *two;
4341
4342 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4343 return;
4344
4345 /* This may look odd, but it is a preparation for
4346 * feeding "there are unchanged files which should
4347 * not produce diffs, but when you are doing copy
4348 * detection you would need them, so here they are"
4349 * entries to the diff-core. They will be prefixed
4350 * with something like '=' or '*' (I haven't decided
4351 * which but should not make any difference).
4352 * Feeding the same new and old to diff_change()
4353 * also has the same effect.
4354 * Before the final output happens, they are pruned after
4355 * merged into rename/copy pairs as appropriate.
4356 */
4357 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4358 addremove = (addremove == '+' ? '-' :
4359 addremove == '-' ? '+' : addremove);
4360
4361 if (options->prefix &&
4362 strncmp(concatpath, options->prefix, options->prefix_length))
4363 return;
4364
4365 one = alloc_filespec(concatpath);
4366 two = alloc_filespec(concatpath);
4367
4368 if (addremove != '+')
4369 fill_filespec(one, sha1, mode);
4370 if (addremove != '-') {
4371 fill_filespec(two, sha1, mode);
4372 two->dirty_submodule = dirty_submodule;
4373 }
4374
4375 diff_queue(&diff_queued_diff, one, two);
4376 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4377 DIFF_OPT_SET(options, HAS_CHANGES);
4378 }
4379
4380 void diff_change(struct diff_options *options,
4381 unsigned old_mode, unsigned new_mode,
4382 const unsigned char *old_sha1,
4383 const unsigned char *new_sha1,
4384 const char *concatpath,
4385 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4386 {
4387 struct diff_filespec *one, *two;
4388
4389 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4390 is_submodule_ignored(concatpath, options))
4391 return;
4392
4393 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4394 unsigned tmp;
4395 const unsigned char *tmp_c;
4396 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4397 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4398 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4399 new_dirty_submodule = tmp;
4400 }
4401
4402 if (options->prefix &&
4403 strncmp(concatpath, options->prefix, options->prefix_length))
4404 return;
4405
4406 one = alloc_filespec(concatpath);
4407 two = alloc_filespec(concatpath);
4408 fill_filespec(one, old_sha1, old_mode);
4409 fill_filespec(two, new_sha1, new_mode);
4410 one->dirty_submodule = old_dirty_submodule;
4411 two->dirty_submodule = new_dirty_submodule;
4412
4413 diff_queue(&diff_queued_diff, one, two);
4414 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4415 DIFF_OPT_SET(options, HAS_CHANGES);
4416 }
4417
4418 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4419 {
4420 struct diff_filepair *pair;
4421 struct diff_filespec *one, *two;
4422
4423 if (options->prefix &&
4424 strncmp(path, options->prefix, options->prefix_length))
4425 return NULL;
4426
4427 one = alloc_filespec(path);
4428 two = alloc_filespec(path);
4429 pair = diff_queue(&diff_queued_diff, one, two);
4430 pair->is_unmerged = 1;
4431 return pair;
4432 }
4433
4434 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4435 size_t *outsize)
4436 {
4437 struct diff_tempfile *temp;
4438 const char *argv[3];
4439 const char **arg = argv;
4440 struct child_process child;
4441 struct strbuf buf = STRBUF_INIT;
4442 int err = 0;
4443
4444 temp = prepare_temp_file(spec->path, spec);
4445 *arg++ = pgm;
4446 *arg++ = temp->name;
4447 *arg = NULL;
4448
4449 memset(&child, 0, sizeof(child));
4450 child.use_shell = 1;
4451 child.argv = argv;
4452 child.out = -1;
4453 if (start_command(&child)) {
4454 remove_tempfile();
4455 return NULL;
4456 }
4457
4458 if (strbuf_read(&buf, child.out, 0) < 0)
4459 err = error("error reading from textconv command '%s'", pgm);
4460 close(child.out);
4461
4462 if (finish_command(&child) || err) {
4463 strbuf_release(&buf);
4464 remove_tempfile();
4465 return NULL;
4466 }
4467 remove_tempfile();
4468
4469 return strbuf_detach(&buf, outsize);
4470 }
4471
4472 size_t fill_textconv(struct userdiff_driver *driver,
4473 struct diff_filespec *df,
4474 char **outbuf)
4475 {
4476 size_t size;
4477
4478 if (!driver || !driver->textconv) {
4479 if (!DIFF_FILE_VALID(df)) {
4480 *outbuf = "";
4481 return 0;
4482 }
4483 if (diff_populate_filespec(df, 0))
4484 die("unable to read files to diff");
4485 *outbuf = df->data;
4486 return df->size;
4487 }
4488
4489 if (driver->textconv_cache && df->sha1_valid) {
4490 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4491 &size);
4492 if (*outbuf)
4493 return size;
4494 }
4495
4496 *outbuf = run_textconv(driver->textconv, df, &size);
4497 if (!*outbuf)
4498 die("unable to read files to diff");
4499
4500 if (driver->textconv_cache && df->sha1_valid) {
4501 /* ignore errors, as we might be in a readonly repository */
4502 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4503 size);
4504 /*
4505 * we could save up changes and flush them all at the end,
4506 * but we would need an extra call after all diffing is done.
4507 * Since generating a cache entry is the slow path anyway,
4508 * this extra overhead probably isn't a big deal.
4509 */
4510 notes_cache_write(driver->textconv_cache);
4511 }
4512
4513 return size;
4514 }
git with Jakub Narębski modifications