?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
--dirstat: In case of renames, use target filename instead of source filename
[git/jrn.git] / diff.c
blobabd9cd5f33d40bc236f6e1d2946992cc967f0341
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 200;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static struct diff_options default_diff_options;
35
36 static char diff_colors[][COLOR_MAXLEN] = {
37 GIT_COLOR_RESET,
38 GIT_COLOR_NORMAL, /* PLAIN */
39 GIT_COLOR_BOLD, /* METAINFO */
40 GIT_COLOR_CYAN, /* FRAGINFO */
41 GIT_COLOR_RED, /* OLD */
42 GIT_COLOR_GREEN, /* NEW */
43 GIT_COLOR_YELLOW, /* COMMIT */
44 GIT_COLOR_BG_RED, /* WHITESPACE */
45 GIT_COLOR_NORMAL, /* FUNCINFO */
46 };
47
48 static int parse_diff_color_slot(const char *var, int ofs)
49 {
50 if (!strcasecmp(var+ofs, "plain"))
51 return DIFF_PLAIN;
52 if (!strcasecmp(var+ofs, "meta"))
53 return DIFF_METAINFO;
54 if (!strcasecmp(var+ofs, "frag"))
55 return DIFF_FRAGINFO;
56 if (!strcasecmp(var+ofs, "old"))
57 return DIFF_FILE_OLD;
58 if (!strcasecmp(var+ofs, "new"))
59 return DIFF_FILE_NEW;
60 if (!strcasecmp(var+ofs, "commit"))
61 return DIFF_COMMIT;
62 if (!strcasecmp(var+ofs, "whitespace"))
63 return DIFF_WHITESPACE;
64 if (!strcasecmp(var+ofs, "func"))
65 return DIFF_FUNCINFO;
66 return -1;
67 }
68
69 static int git_config_rename(const char *var, const char *value)
70 {
71 if (!value)
72 return DIFF_DETECT_RENAME;
73 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
74 return DIFF_DETECT_COPY;
75 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
76 }
77
78 /*
79 * These are to give UI layer defaults.
80 * The core-level commands such as git-diff-files should
81 * never be affected by the setting of diff.renames
82 * the user happens to have in the configuration file.
83 */
84 int git_diff_ui_config(const char *var, const char *value, void *cb)
85 {
86 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
87 diff_use_color_default = git_config_colorbool(var, value, -1);
88 return 0;
89 }
90 if (!strcmp(var, "diff.renames")) {
91 diff_detect_rename_default = git_config_rename(var, value);
92 return 0;
93 }
94 if (!strcmp(var, "diff.autorefreshindex")) {
95 diff_auto_refresh_index = git_config_bool(var, value);
96 return 0;
97 }
98 if (!strcmp(var, "diff.mnemonicprefix")) {
99 diff_mnemonic_prefix = git_config_bool(var, value);
100 return 0;
101 }
102 if (!strcmp(var, "diff.noprefix")) {
103 diff_no_prefix = git_config_bool(var, value);
104 return 0;
105 }
106 if (!strcmp(var, "diff.external"))
107 return git_config_string(&external_diff_cmd_cfg, var, value);
108 if (!strcmp(var, "diff.wordregex"))
109 return git_config_string(&diff_word_regex_cfg, var, value);
110
111 if (!strcmp(var, "diff.ignoresubmodules"))
112 handle_ignore_submodules_arg(&default_diff_options, value);
113
114 return git_diff_basic_config(var, value, cb);
115 }
116
117 int git_diff_basic_config(const char *var, const char *value, void *cb)
118 {
119 if (!strcmp(var, "diff.renamelimit")) {
120 diff_rename_limit_default = git_config_int(var, value);
121 return 0;
122 }
123
124 switch (userdiff_config(var, value)) {
125 case 0: break;
126 case -1: return -1;
127 default: return 0;
128 }
129
130 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
131 int slot = parse_diff_color_slot(var, 11);
132 if (slot < 0)
133 return 0;
134 if (!value)
135 return config_error_nonbool(var);
136 color_parse(value, var, diff_colors[slot]);
137 return 0;
138 }
139
140 /* like GNU diff's --suppress-blank-empty option */
141 if (!strcmp(var, "diff.suppressblankempty") ||
142 /* for backwards compatibility */
143 !strcmp(var, "diff.suppress-blank-empty")) {
144 diff_suppress_blank_empty = git_config_bool(var, value);
145 return 0;
146 }
147
148 if (!prefixcmp(var, "submodule."))
149 return parse_submodule_config_option(var, value);
150
151 return git_color_default_config(var, value, cb);
152 }
153
154 static char *quote_two(const char *one, const char *two)
155 {
156 int need_one = quote_c_style(one, NULL, NULL, 1);
157 int need_two = quote_c_style(two, NULL, NULL, 1);
158 struct strbuf res = STRBUF_INIT;
159
160 if (need_one + need_two) {
161 strbuf_addch(&res, '"');
162 quote_c_style(one, &res, NULL, 1);
163 quote_c_style(two, &res, NULL, 1);
164 strbuf_addch(&res, '"');
165 } else {
166 strbuf_addstr(&res, one);
167 strbuf_addstr(&res, two);
168 }
169 return strbuf_detach(&res, NULL);
170 }
171
172 static const char *external_diff(void)
173 {
174 static const char *external_diff_cmd = NULL;
175 static int done_preparing = 0;
176
177 if (done_preparing)
178 return external_diff_cmd;
179 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
180 if (!external_diff_cmd)
181 external_diff_cmd = external_diff_cmd_cfg;
182 done_preparing = 1;
183 return external_diff_cmd;
184 }
185
186 static struct diff_tempfile {
187 const char *name; /* filename external diff should read from */
188 char hex[41];
189 char mode[10];
190 char tmp_path[PATH_MAX];
191 } diff_temp[2];
192
193 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
194
195 struct emit_callback {
196 int color_diff;
197 unsigned ws_rule;
198 int blank_at_eof_in_preimage;
199 int blank_at_eof_in_postimage;
200 int lno_in_preimage;
201 int lno_in_postimage;
202 sane_truncate_fn truncate;
203 const char **label_path;
204 struct diff_words_data *diff_words;
205 struct diff_options *opt;
206 int *found_changesp;
207 struct strbuf *header;
208 };
209
210 static int count_lines(const char *data, int size)
211 {
212 int count, ch, completely_empty = 1, nl_just_seen = 0;
213 count = 0;
214 while (0 < size--) {
215 ch = *data++;
216 if (ch == '\n') {
217 count++;
218 nl_just_seen = 1;
219 completely_empty = 0;
220 }
221 else {
222 nl_just_seen = 0;
223 completely_empty = 0;
224 }
225 }
226 if (completely_empty)
227 return 0;
228 if (!nl_just_seen)
229 count++; /* no trailing newline */
230 return count;
231 }
232
233 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
234 {
235 if (!DIFF_FILE_VALID(one)) {
236 mf->ptr = (char *)""; /* does not matter */
237 mf->size = 0;
238 return 0;
239 }
240 else if (diff_populate_filespec(one, 0))
241 return -1;
242
243 mf->ptr = one->data;
244 mf->size = one->size;
245 return 0;
246 }
247
248 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
249 static unsigned long diff_filespec_size(struct diff_filespec *one)
250 {
251 if (!DIFF_FILE_VALID(one))
252 return 0;
253 diff_populate_filespec(one, 1);
254 return one->size;
255 }
256
257 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
258 {
259 char *ptr = mf->ptr;
260 long size = mf->size;
261 int cnt = 0;
262
263 if (!size)
264 return cnt;
265 ptr += size - 1; /* pointing at the very end */
266 if (*ptr != '\n')
267 ; /* incomplete line */
268 else
269 ptr--; /* skip the last LF */
270 while (mf->ptr < ptr) {
271 char *prev_eol;
272 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
273 if (*prev_eol == '\n')
274 break;
275 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
276 break;
277 cnt++;
278 ptr = prev_eol - 1;
279 }
280 return cnt;
281 }
282
283 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
284 struct emit_callback *ecbdata)
285 {
286 int l1, l2, at;
287 unsigned ws_rule = ecbdata->ws_rule;
288 l1 = count_trailing_blank(mf1, ws_rule);
289 l2 = count_trailing_blank(mf2, ws_rule);
290 if (l2 <= l1) {
291 ecbdata->blank_at_eof_in_preimage = 0;
292 ecbdata->blank_at_eof_in_postimage = 0;
293 return;
294 }
295 at = count_lines(mf1->ptr, mf1->size);
296 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
297
298 at = count_lines(mf2->ptr, mf2->size);
299 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
300 }
301
302 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
303 int first, const char *line, int len)
304 {
305 int has_trailing_newline, has_trailing_carriage_return;
306 int nofirst;
307 FILE *file = o->file;
308
309 if (o->output_prefix) {
310 struct strbuf *msg = NULL;
311 msg = o->output_prefix(o, o->output_prefix_data);
312 assert(msg);
313 fwrite(msg->buf, msg->len, 1, file);
314 }
315
316 if (len == 0) {
317 has_trailing_newline = (first == '\n');
318 has_trailing_carriage_return = (!has_trailing_newline &&
319 (first == '\r'));
320 nofirst = has_trailing_newline || has_trailing_carriage_return;
321 } else {
322 has_trailing_newline = (len > 0 && line[len-1] == '\n');
323 if (has_trailing_newline)
324 len--;
325 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
326 if (has_trailing_carriage_return)
327 len--;
328 nofirst = 0;
329 }
330
331 if (len || !nofirst) {
332 fputs(set, file);
333 if (!nofirst)
334 fputc(first, file);
335 fwrite(line, len, 1, file);
336 fputs(reset, file);
337 }
338 if (has_trailing_carriage_return)
339 fputc('\r', file);
340 if (has_trailing_newline)
341 fputc('\n', file);
342 }
343
344 static void emit_line(struct diff_options *o, const char *set, const char *reset,
345 const char *line, int len)
346 {
347 emit_line_0(o, set, reset, line[0], line+1, len-1);
348 }
349
350 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
351 {
352 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
353 ecbdata->blank_at_eof_in_preimage &&
354 ecbdata->blank_at_eof_in_postimage &&
355 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
356 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
357 return 0;
358 return ws_blank_line(line, len, ecbdata->ws_rule);
359 }
360
361 static void emit_add_line(const char *reset,
362 struct emit_callback *ecbdata,
363 const char *line, int len)
364 {
365 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
366 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
367
368 if (!*ws)
369 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
370 else if (new_blank_line_at_eof(ecbdata, line, len))
371 /* Blank line at EOF - paint '+' as well */
372 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
373 else {
374 /* Emit just the prefix, then the rest. */
375 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
376 ws_check_emit(line, len, ecbdata->ws_rule,
377 ecbdata->opt->file, set, reset, ws);
378 }
379 }
380
381 static void emit_hunk_header(struct emit_callback *ecbdata,
382 const char *line, int len)
383 {
384 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
385 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
386 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
387 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
388 static const char atat[2] = { '@', '@' };
389 const char *cp, *ep;
390 struct strbuf msgbuf = STRBUF_INIT;
391 int org_len = len;
392 int i = 1;
393
394 /*
395 * As a hunk header must begin with "@@ -<old>, +<new> @@",
396 * it always is at least 10 bytes long.
397 */
398 if (len < 10 ||
399 memcmp(line, atat, 2) ||
400 !(ep = memmem(line + 2, len - 2, atat, 2))) {
401 emit_line(ecbdata->opt, plain, reset, line, len);
402 return;
403 }
404 ep += 2; /* skip over @@ */
405
406 /* The hunk header in fraginfo color */
407 strbuf_add(&msgbuf, frag, strlen(frag));
408 strbuf_add(&msgbuf, line, ep - line);
409 strbuf_add(&msgbuf, reset, strlen(reset));
410
411 /*
412 * trailing "\r\n"
413 */
414 for ( ; i < 3; i++)
415 if (line[len - i] == '\r' || line[len - i] == '\n')
416 len--;
417
418 /* blank before the func header */
419 for (cp = ep; ep - line < len; ep++)
420 if (*ep != ' ' && *ep != '\t')
421 break;
422 if (ep != cp) {
423 strbuf_add(&msgbuf, plain, strlen(plain));
424 strbuf_add(&msgbuf, cp, ep - cp);
425 strbuf_add(&msgbuf, reset, strlen(reset));
426 }
427
428 if (ep < line + len) {
429 strbuf_add(&msgbuf, func, strlen(func));
430 strbuf_add(&msgbuf, ep, line + len - ep);
431 strbuf_add(&msgbuf, reset, strlen(reset));
432 }
433
434 strbuf_add(&msgbuf, line + len, org_len - len);
435 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
436 strbuf_release(&msgbuf);
437 }
438
439 static struct diff_tempfile *claim_diff_tempfile(void) {
440 int i;
441 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
442 if (!diff_temp[i].name)
443 return diff_temp + i;
444 die("BUG: diff is failing to clean up its tempfiles");
445 }
446
447 static int remove_tempfile_installed;
448
449 static void remove_tempfile(void)
450 {
451 int i;
452 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
453 if (diff_temp[i].name == diff_temp[i].tmp_path)
454 unlink_or_warn(diff_temp[i].name);
455 diff_temp[i].name = NULL;
456 }
457 }
458
459 static void remove_tempfile_on_signal(int signo)
460 {
461 remove_tempfile();
462 sigchain_pop(signo);
463 raise(signo);
464 }
465
466 static void print_line_count(FILE *file, int count)
467 {
468 switch (count) {
469 case 0:
470 fprintf(file, "0,0");
471 break;
472 case 1:
473 fprintf(file, "1");
474 break;
475 default:
476 fprintf(file, "1,%d", count);
477 break;
478 }
479 }
480
481 static void emit_rewrite_lines(struct emit_callback *ecb,
482 int prefix, const char *data, int size)
483 {
484 const char *endp = NULL;
485 static const char *nneof = " No newline at end of file\n";
486 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
487 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
488
489 while (0 < size) {
490 int len;
491
492 endp = memchr(data, '\n', size);
493 len = endp ? (endp - data + 1) : size;
494 if (prefix != '+') {
495 ecb->lno_in_preimage++;
496 emit_line_0(ecb->opt, old, reset, '-',
497 data, len);
498 } else {
499 ecb->lno_in_postimage++;
500 emit_add_line(reset, ecb, data, len);
501 }
502 size -= len;
503 data += len;
504 }
505 if (!endp) {
506 const char *plain = diff_get_color(ecb->color_diff,
507 DIFF_PLAIN);
508 emit_line_0(ecb->opt, plain, reset, '\\',
509 nneof, strlen(nneof));
510 }
511 }
512
513 static void emit_rewrite_diff(const char *name_a,
514 const char *name_b,
515 struct diff_filespec *one,
516 struct diff_filespec *two,
517 struct userdiff_driver *textconv_one,
518 struct userdiff_driver *textconv_two,
519 struct diff_options *o)
520 {
521 int lc_a, lc_b;
522 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
523 const char *name_a_tab, *name_b_tab;
524 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
525 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
526 const char *reset = diff_get_color(color_diff, DIFF_RESET);
527 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
528 const char *a_prefix, *b_prefix;
529 char *data_one, *data_two;
530 size_t size_one, size_two;
531 struct emit_callback ecbdata;
532 char *line_prefix = "";
533 struct strbuf *msgbuf;
534
535 if (o && o->output_prefix) {
536 msgbuf = o->output_prefix(o, o->output_prefix_data);
537 line_prefix = msgbuf->buf;
538 }
539
540 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
541 a_prefix = o->b_prefix;
542 b_prefix = o->a_prefix;
543 } else {
544 a_prefix = o->a_prefix;
545 b_prefix = o->b_prefix;
546 }
547
548 name_a += (*name_a == '/');
549 name_b += (*name_b == '/');
550 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
551 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
552
553 strbuf_reset(&a_name);
554 strbuf_reset(&b_name);
555 quote_two_c_style(&a_name, a_prefix, name_a, 0);
556 quote_two_c_style(&b_name, b_prefix, name_b, 0);
557
558 size_one = fill_textconv(textconv_one, one, &data_one);
559 size_two = fill_textconv(textconv_two, two, &data_two);
560
561 memset(&ecbdata, 0, sizeof(ecbdata));
562 ecbdata.color_diff = color_diff;
563 ecbdata.found_changesp = &o->found_changes;
564 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
565 ecbdata.opt = o;
566 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
567 mmfile_t mf1, mf2;
568 mf1.ptr = (char *)data_one;
569 mf2.ptr = (char *)data_two;
570 mf1.size = size_one;
571 mf2.size = size_two;
572 check_blank_at_eof(&mf1, &mf2, &ecbdata);
573 }
574 ecbdata.lno_in_preimage = 1;
575 ecbdata.lno_in_postimage = 1;
576
577 lc_a = count_lines(data_one, size_one);
578 lc_b = count_lines(data_two, size_two);
579 fprintf(o->file,
580 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
581 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
582 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
583 line_prefix, fraginfo);
584 print_line_count(o->file, lc_a);
585 fprintf(o->file, " +");
586 print_line_count(o->file, lc_b);
587 fprintf(o->file, " @@%s\n", reset);
588 if (lc_a)
589 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
590 if (lc_b)
591 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
592 if (textconv_one)
593 free((char *)data_one);
594 if (textconv_two)
595 free((char *)data_two);
596 }
597
598 struct diff_words_buffer {
599 mmfile_t text;
600 long alloc;
601 struct diff_words_orig {
602 const char *begin, *end;
603 } *orig;
604 int orig_nr, orig_alloc;
605 };
606
607 static void diff_words_append(char *line, unsigned long len,
608 struct diff_words_buffer *buffer)
609 {
610 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
611 line++;
612 len--;
613 memcpy(buffer->text.ptr + buffer->text.size, line, len);
614 buffer->text.size += len;
615 buffer->text.ptr[buffer->text.size] = '\0';
616 }
617
618 struct diff_words_style_elem {
619 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 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 /* don't print the prefix character */
1054 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1055 }
1056 return;
1057 }
1058
1059 if (line[0] != '+') {
1060 const char *color =
1061 diff_get_color(ecbdata->color_diff,
1062 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1063 ecbdata->lno_in_preimage++;
1064 if (line[0] == ' ')
1065 ecbdata->lno_in_postimage++;
1066 emit_line(ecbdata->opt, color, reset, line, len);
1067 } else {
1068 ecbdata->lno_in_postimage++;
1069 emit_add_line(reset, ecbdata, line + 1, len - 1);
1070 }
1071 }
1072
1073 static char *pprint_rename(const char *a, const char *b)
1074 {
1075 const char *old = a;
1076 const char *new = b;
1077 struct strbuf name = STRBUF_INIT;
1078 int pfx_length, sfx_length;
1079 int len_a = strlen(a);
1080 int len_b = strlen(b);
1081 int a_midlen, b_midlen;
1082 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1083 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1084
1085 if (qlen_a || qlen_b) {
1086 quote_c_style(a, &name, NULL, 0);
1087 strbuf_addstr(&name, " => ");
1088 quote_c_style(b, &name, NULL, 0);
1089 return strbuf_detach(&name, NULL);
1090 }
1091
1092 /* Find common prefix */
1093 pfx_length = 0;
1094 while (*old && *new && *old == *new) {
1095 if (*old == '/')
1096 pfx_length = old - a + 1;
1097 old++;
1098 new++;
1099 }
1100
1101 /* Find common suffix */
1102 old = a + len_a;
1103 new = b + len_b;
1104 sfx_length = 0;
1105 while (a <= old && b <= new && *old == *new) {
1106 if (*old == '/')
1107 sfx_length = len_a - (old - a);
1108 old--;
1109 new--;
1110 }
1111
1112 /*
1113 * pfx{mid-a => mid-b}sfx
1114 * {pfx-a => pfx-b}sfx
1115 * pfx{sfx-a => sfx-b}
1116 * name-a => name-b
1117 */
1118 a_midlen = len_a - pfx_length - sfx_length;
1119 b_midlen = len_b - pfx_length - sfx_length;
1120 if (a_midlen < 0)
1121 a_midlen = 0;
1122 if (b_midlen < 0)
1123 b_midlen = 0;
1124
1125 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1126 if (pfx_length + sfx_length) {
1127 strbuf_add(&name, a, pfx_length);
1128 strbuf_addch(&name, '{');
1129 }
1130 strbuf_add(&name, a + pfx_length, a_midlen);
1131 strbuf_addstr(&name, " => ");
1132 strbuf_add(&name, b + pfx_length, b_midlen);
1133 if (pfx_length + sfx_length) {
1134 strbuf_addch(&name, '}');
1135 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1136 }
1137 return strbuf_detach(&name, NULL);
1138 }
1139
1140 struct diffstat_t {
1141 int nr;
1142 int alloc;
1143 struct diffstat_file {
1144 char *from_name;
1145 char *name;
1146 char *print_name;
1147 unsigned is_unmerged:1;
1148 unsigned is_binary:1;
1149 unsigned is_renamed:1;
1150 uintmax_t added, deleted;
1151 } **files;
1152 };
1153
1154 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1155 const char *name_a,
1156 const char *name_b)
1157 {
1158 struct diffstat_file *x;
1159 x = xcalloc(sizeof (*x), 1);
1160 if (diffstat->nr == diffstat->alloc) {
1161 diffstat->alloc = alloc_nr(diffstat->alloc);
1162 diffstat->files = xrealloc(diffstat->files,
1163 diffstat->alloc * sizeof(x));
1164 }
1165 diffstat->files[diffstat->nr++] = x;
1166 if (name_b) {
1167 x->from_name = xstrdup(name_a);
1168 x->name = xstrdup(name_b);
1169 x->is_renamed = 1;
1170 }
1171 else {
1172 x->from_name = NULL;
1173 x->name = xstrdup(name_a);
1174 }
1175 return x;
1176 }
1177
1178 static void diffstat_consume(void *priv, char *line, unsigned long len)
1179 {
1180 struct diffstat_t *diffstat = priv;
1181 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1182
1183 if (line[0] == '+')
1184 x->added++;
1185 else if (line[0] == '-')
1186 x->deleted++;
1187 }
1188
1189 const char mime_boundary_leader[] = "------------";
1190
1191 static int scale_linear(int it, int width, int max_change)
1192 {
1193 /*
1194 * make sure that at least one '-' is printed if there were deletions,
1195 * and likewise for '+'.
1196 */
1197 if (max_change < 2)
1198 return it;
1199 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1200 }
1201
1202 static void show_name(FILE *file,
1203 const char *prefix, const char *name, int len)
1204 {
1205 fprintf(file, " %s%-*s |", prefix, len, name);
1206 }
1207
1208 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1209 {
1210 if (cnt <= 0)
1211 return;
1212 fprintf(file, "%s", set);
1213 while (cnt--)
1214 putc(ch, file);
1215 fprintf(file, "%s", reset);
1216 }
1217
1218 static void fill_print_name(struct diffstat_file *file)
1219 {
1220 char *pname;
1221
1222 if (file->print_name)
1223 return;
1224
1225 if (!file->is_renamed) {
1226 struct strbuf buf = STRBUF_INIT;
1227 if (quote_c_style(file->name, &buf, NULL, 0)) {
1228 pname = strbuf_detach(&buf, NULL);
1229 } else {
1230 pname = file->name;
1231 strbuf_release(&buf);
1232 }
1233 } else {
1234 pname = pprint_rename(file->from_name, file->name);
1235 }
1236 file->print_name = pname;
1237 }
1238
1239 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1240 {
1241 int i, len, add, del, adds = 0, dels = 0;
1242 uintmax_t max_change = 0, max_len = 0;
1243 int total_files = data->nr;
1244 int width, name_width;
1245 const char *reset, *set, *add_c, *del_c;
1246 const char *line_prefix = "";
1247 struct strbuf *msg = NULL;
1248
1249 if (data->nr == 0)
1250 return;
1251
1252 if (options->output_prefix) {
1253 msg = options->output_prefix(options, options->output_prefix_data);
1254 line_prefix = msg->buf;
1255 }
1256
1257 width = options->stat_width ? options->stat_width : 80;
1258 name_width = options->stat_name_width ? options->stat_name_width : 50;
1259
1260 /* Sanity: give at least 5 columns to the graph,
1261 * but leave at least 10 columns for the name.
1262 */
1263 if (width < 25)
1264 width = 25;
1265 if (name_width < 10)
1266 name_width = 10;
1267 else if (width < name_width + 15)
1268 name_width = width - 15;
1269
1270 /* Find the longest filename and max number of changes */
1271 reset = diff_get_color_opt(options, DIFF_RESET);
1272 set = diff_get_color_opt(options, DIFF_PLAIN);
1273 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1274 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1275
1276 for (i = 0; i < data->nr; i++) {
1277 struct diffstat_file *file = data->files[i];
1278 uintmax_t change = file->added + file->deleted;
1279 fill_print_name(file);
1280 len = strlen(file->print_name);
1281 if (max_len < len)
1282 max_len = len;
1283
1284 if (file->is_binary || file->is_unmerged)
1285 continue;
1286 if (max_change < change)
1287 max_change = change;
1288 }
1289
1290 /* Compute the width of the graph part;
1291 * 10 is for one blank at the beginning of the line plus
1292 * " | count " between the name and the graph.
1293 *
1294 * From here on, name_width is the width of the name area,
1295 * and width is the width of the graph area.
1296 */
1297 name_width = (name_width < max_len) ? name_width : max_len;
1298 if (width < (name_width + 10) + max_change)
1299 width = width - (name_width + 10);
1300 else
1301 width = max_change;
1302
1303 for (i = 0; i < data->nr; i++) {
1304 const char *prefix = "";
1305 char *name = data->files[i]->print_name;
1306 uintmax_t added = data->files[i]->added;
1307 uintmax_t deleted = data->files[i]->deleted;
1308 int name_len;
1309
1310 /*
1311 * "scale" the filename
1312 */
1313 len = name_width;
1314 name_len = strlen(name);
1315 if (name_width < name_len) {
1316 char *slash;
1317 prefix = "...";
1318 len -= 3;
1319 name += name_len - len;
1320 slash = strchr(name, '/');
1321 if (slash)
1322 name = slash;
1323 }
1324
1325 if (data->files[i]->is_binary) {
1326 fprintf(options->file, "%s", line_prefix);
1327 show_name(options->file, prefix, name, len);
1328 fprintf(options->file, " Bin ");
1329 fprintf(options->file, "%s%"PRIuMAX"%s",
1330 del_c, deleted, reset);
1331 fprintf(options->file, " -> ");
1332 fprintf(options->file, "%s%"PRIuMAX"%s",
1333 add_c, added, reset);
1334 fprintf(options->file, " bytes");
1335 fprintf(options->file, "\n");
1336 continue;
1337 }
1338 else if (data->files[i]->is_unmerged) {
1339 fprintf(options->file, "%s", line_prefix);
1340 show_name(options->file, prefix, name, len);
1341 fprintf(options->file, " Unmerged\n");
1342 continue;
1343 }
1344 else if (!data->files[i]->is_renamed &&
1345 (added + deleted == 0)) {
1346 total_files--;
1347 continue;
1348 }
1349
1350 /*
1351 * scale the add/delete
1352 */
1353 add = added;
1354 del = deleted;
1355 adds += add;
1356 dels += del;
1357
1358 if (width <= max_change) {
1359 add = scale_linear(add, width, max_change);
1360 del = scale_linear(del, width, max_change);
1361 }
1362 fprintf(options->file, "%s", line_prefix);
1363 show_name(options->file, prefix, name, len);
1364 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1365 added + deleted ? " " : "");
1366 show_graph(options->file, '+', add, add_c, reset);
1367 show_graph(options->file, '-', del, del_c, reset);
1368 fprintf(options->file, "\n");
1369 }
1370 fprintf(options->file, "%s", line_prefix);
1371 fprintf(options->file,
1372 " %d files changed, %d insertions(+), %d deletions(-)\n",
1373 total_files, adds, dels);
1374 }
1375
1376 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1377 {
1378 int i, adds = 0, dels = 0, total_files = data->nr;
1379
1380 if (data->nr == 0)
1381 return;
1382
1383 for (i = 0; i < data->nr; i++) {
1384 if (!data->files[i]->is_binary &&
1385 !data->files[i]->is_unmerged) {
1386 int added = data->files[i]->added;
1387 int deleted= data->files[i]->deleted;
1388 if (!data->files[i]->is_renamed &&
1389 (added + deleted == 0)) {
1390 total_files--;
1391 } else {
1392 adds += added;
1393 dels += deleted;
1394 }
1395 }
1396 }
1397 if (options->output_prefix) {
1398 struct strbuf *msg = NULL;
1399 msg = options->output_prefix(options,
1400 options->output_prefix_data);
1401 fprintf(options->file, "%s", msg->buf);
1402 }
1403 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1404 total_files, adds, dels);
1405 }
1406
1407 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1408 {
1409 int i;
1410
1411 if (data->nr == 0)
1412 return;
1413
1414 for (i = 0; i < data->nr; i++) {
1415 struct diffstat_file *file = data->files[i];
1416
1417 if (options->output_prefix) {
1418 struct strbuf *msg = NULL;
1419 msg = options->output_prefix(options,
1420 options->output_prefix_data);
1421 fprintf(options->file, "%s", msg->buf);
1422 }
1423
1424 if (file->is_binary)
1425 fprintf(options->file, "-\t-\t");
1426 else
1427 fprintf(options->file,
1428 "%"PRIuMAX"\t%"PRIuMAX"\t",
1429 file->added, file->deleted);
1430 if (options->line_termination) {
1431 fill_print_name(file);
1432 if (!file->is_renamed)
1433 write_name_quoted(file->name, options->file,
1434 options->line_termination);
1435 else {
1436 fputs(file->print_name, options->file);
1437 putc(options->line_termination, options->file);
1438 }
1439 } else {
1440 if (file->is_renamed) {
1441 putc('\0', options->file);
1442 write_name_quoted(file->from_name, options->file, '\0');
1443 }
1444 write_name_quoted(file->name, options->file, '\0');
1445 }
1446 }
1447 }
1448
1449 struct dirstat_file {
1450 const char *name;
1451 unsigned long changed;
1452 };
1453
1454 struct dirstat_dir {
1455 struct dirstat_file *files;
1456 int alloc, nr, percent, cumulative;
1457 };
1458
1459 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1460 unsigned long changed, const char *base, int baselen)
1461 {
1462 unsigned long this_dir = 0;
1463 unsigned int sources = 0;
1464 const char *line_prefix = "";
1465 struct strbuf *msg = NULL;
1466
1467 if (opt->output_prefix) {
1468 msg = opt->output_prefix(opt, opt->output_prefix_data);
1469 line_prefix = msg->buf;
1470 }
1471
1472 while (dir->nr) {
1473 struct dirstat_file *f = dir->files;
1474 int namelen = strlen(f->name);
1475 unsigned long this;
1476 char *slash;
1477
1478 if (namelen < baselen)
1479 break;
1480 if (memcmp(f->name, base, baselen))
1481 break;
1482 slash = strchr(f->name + baselen, '/');
1483 if (slash) {
1484 int newbaselen = slash + 1 - f->name;
1485 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1486 sources++;
1487 } else {
1488 this = f->changed;
1489 dir->files++;
1490 dir->nr--;
1491 sources += 2;
1492 }
1493 this_dir += this;
1494 }
1495
1496 /*
1497 * We don't report dirstat's for
1498 * - the top level
1499 * - or cases where everything came from a single directory
1500 * under this directory (sources == 1).
1501 */
1502 if (baselen && sources != 1) {
1503 int permille = this_dir * 1000 / changed;
1504 if (permille) {
1505 int percent = permille / 10;
1506 if (percent >= dir->percent) {
1507 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1508 percent, permille % 10, baselen, base);
1509 if (!dir->cumulative)
1510 return 0;
1511 }
1512 }
1513 }
1514 return this_dir;
1515 }
1516
1517 static int dirstat_compare(const void *_a, const void *_b)
1518 {
1519 const struct dirstat_file *a = _a;
1520 const struct dirstat_file *b = _b;
1521 return strcmp(a->name, b->name);
1522 }
1523
1524 static void show_dirstat(struct diff_options *options)
1525 {
1526 int i;
1527 unsigned long changed;
1528 struct dirstat_dir dir;
1529 struct diff_queue_struct *q = &diff_queued_diff;
1530
1531 dir.files = NULL;
1532 dir.alloc = 0;
1533 dir.nr = 0;
1534 dir.percent = options->dirstat_percent;
1535 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1536
1537 changed = 0;
1538 for (i = 0; i < q->nr; i++) {
1539 struct diff_filepair *p = q->queue[i];
1540 const char *name;
1541 unsigned long copied, added, damage;
1542 int content_changed;
1543
1544 name = p->two->path ? p->two->path : p->one->path;
1545
1546 if (p->one->sha1_valid && p->two->sha1_valid)
1547 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1548 else
1549 content_changed = 1;
1550
1551 if (!content_changed) {
1552 /*
1553 * The SHA1 has not changed, so pre-/post-content is
1554 * identical. We can therefore skip looking at the
1555 * file contents altogether.
1556 */
1557 damage = 0;
1558 goto found_damage;
1559 }
1560
1561 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1562 /*
1563 * In --dirstat-by-file mode, we don't really need to
1564 * look at the actual file contents at all.
1565 * The fact that the SHA1 changed is enough for us to
1566 * add this file to the list of results
1567 * (with each file contributing equal damage).
1568 */
1569 damage = 1;
1570 goto found_damage;
1571 }
1572
1573 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1574 diff_populate_filespec(p->one, 0);
1575 diff_populate_filespec(p->two, 0);
1576 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1577 &copied, &added);
1578 diff_free_filespec_data(p->one);
1579 diff_free_filespec_data(p->two);
1580 } else if (DIFF_FILE_VALID(p->one)) {
1581 diff_populate_filespec(p->one, 1);
1582 copied = added = 0;
1583 diff_free_filespec_data(p->one);
1584 } else if (DIFF_FILE_VALID(p->two)) {
1585 diff_populate_filespec(p->two, 1);
1586 copied = 0;
1587 added = p->two->size;
1588 diff_free_filespec_data(p->two);
1589 } else
1590 continue;
1591
1592 /*
1593 * Original minus copied is the removed material,
1594 * added is the new material. They are both damages
1595 * made to the preimage.
1596 * If the resulting damage is zero, we know that
1597 * diffcore_count_changes() considers the two entries to
1598 * be identical, but since content_changed is true, we
1599 * know that there must have been _some_ kind of change,
1600 * so we force all entries to have damage > 0.
1601 */
1602 damage = (p->one->size - copied) + added;
1603 if (!damage)
1604 damage = 1;
1605
1606 found_damage:
1607 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1608 dir.files[dir.nr].name = name;
1609 dir.files[dir.nr].changed = damage;
1610 changed += damage;
1611 dir.nr++;
1612 }
1613
1614 /* This can happen even with many files, if everything was renames */
1615 if (!changed)
1616 return;
1617
1618 /* Show all directories with more than x% of the changes */
1619 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1620 gather_dirstat(options, &dir, changed, "", 0);
1621 }
1622
1623 static void free_diffstat_info(struct diffstat_t *diffstat)
1624 {
1625 int i;
1626 for (i = 0; i < diffstat->nr; i++) {
1627 struct diffstat_file *f = diffstat->files[i];
1628 if (f->name != f->print_name)
1629 free(f->print_name);
1630 free(f->name);
1631 free(f->from_name);
1632 free(f);
1633 }
1634 free(diffstat->files);
1635 }
1636
1637 struct checkdiff_t {
1638 const char *filename;
1639 int lineno;
1640 int conflict_marker_size;
1641 struct diff_options *o;
1642 unsigned ws_rule;
1643 unsigned status;
1644 };
1645
1646 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1647 {
1648 char firstchar;
1649 int cnt;
1650
1651 if (len < marker_size + 1)
1652 return 0;
1653 firstchar = line[0];
1654 switch (firstchar) {
1655 case '=': case '>': case '<': case '|':
1656 break;
1657 default:
1658 return 0;
1659 }
1660 for (cnt = 1; cnt < marker_size; cnt++)
1661 if (line[cnt] != firstchar)
1662 return 0;
1663 /* line[1] thru line[marker_size-1] are same as firstchar */
1664 if (len < marker_size + 1 || !isspace(line[marker_size]))
1665 return 0;
1666 return 1;
1667 }
1668
1669 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1670 {
1671 struct checkdiff_t *data = priv;
1672 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1673 int marker_size = data->conflict_marker_size;
1674 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1675 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1676 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1677 char *err;
1678 char *line_prefix = "";
1679 struct strbuf *msgbuf;
1680
1681 assert(data->o);
1682 if (data->o->output_prefix) {
1683 msgbuf = data->o->output_prefix(data->o,
1684 data->o->output_prefix_data);
1685 line_prefix = msgbuf->buf;
1686 }
1687
1688 if (line[0] == '+') {
1689 unsigned bad;
1690 data->lineno++;
1691 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1692 data->status |= 1;
1693 fprintf(data->o->file,
1694 "%s%s:%d: leftover conflict marker\n",
1695 line_prefix, data->filename, data->lineno);
1696 }
1697 bad = ws_check(line + 1, len - 1, data->ws_rule);
1698 if (!bad)
1699 return;
1700 data->status |= bad;
1701 err = whitespace_error_string(bad);
1702 fprintf(data->o->file, "%s%s:%d: %s.\n",
1703 line_prefix, data->filename, data->lineno, err);
1704 free(err);
1705 emit_line(data->o, set, reset, line, 1);
1706 ws_check_emit(line + 1, len - 1, data->ws_rule,
1707 data->o->file, set, reset, ws);
1708 } else if (line[0] == ' ') {
1709 data->lineno++;
1710 } else if (line[0] == '@') {
1711 char *plus = strchr(line, '+');
1712 if (plus)
1713 data->lineno = strtol(plus, NULL, 10) - 1;
1714 else
1715 die("invalid diff");
1716 }
1717 }
1718
1719 static unsigned char *deflate_it(char *data,
1720 unsigned long size,
1721 unsigned long *result_size)
1722 {
1723 int bound;
1724 unsigned char *deflated;
1725 z_stream stream;
1726
1727 memset(&stream, 0, sizeof(stream));
1728 deflateInit(&stream, zlib_compression_level);
1729 bound = deflateBound(&stream, size);
1730 deflated = xmalloc(bound);
1731 stream.next_out = deflated;
1732 stream.avail_out = bound;
1733
1734 stream.next_in = (unsigned char *)data;
1735 stream.avail_in = size;
1736 while (deflate(&stream, Z_FINISH) == Z_OK)
1737 ; /* nothing */
1738 deflateEnd(&stream);
1739 *result_size = stream.total_out;
1740 return deflated;
1741 }
1742
1743 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1744 {
1745 void *cp;
1746 void *delta;
1747 void *deflated;
1748 void *data;
1749 unsigned long orig_size;
1750 unsigned long delta_size;
1751 unsigned long deflate_size;
1752 unsigned long data_size;
1753
1754 /* We could do deflated delta, or we could do just deflated two,
1755 * whichever is smaller.
1756 */
1757 delta = NULL;
1758 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1759 if (one->size && two->size) {
1760 delta = diff_delta(one->ptr, one->size,
1761 two->ptr, two->size,
1762 &delta_size, deflate_size);
1763 if (delta) {
1764 void *to_free = delta;
1765 orig_size = delta_size;
1766 delta = deflate_it(delta, delta_size, &delta_size);
1767 free(to_free);
1768 }
1769 }
1770
1771 if (delta && delta_size < deflate_size) {
1772 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1773 free(deflated);
1774 data = delta;
1775 data_size = delta_size;
1776 }
1777 else {
1778 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1779 free(delta);
1780 data = deflated;
1781 data_size = deflate_size;
1782 }
1783
1784 /* emit data encoded in base85 */
1785 cp = data;
1786 while (data_size) {
1787 int bytes = (52 < data_size) ? 52 : data_size;
1788 char line[70];
1789 data_size -= bytes;
1790 if (bytes <= 26)
1791 line[0] = bytes + 'A' - 1;
1792 else
1793 line[0] = bytes - 26 + 'a' - 1;
1794 encode_85(line + 1, cp, bytes);
1795 cp = (char *) cp + bytes;
1796 fprintf(file, "%s", prefix);
1797 fputs(line, file);
1798 fputc('\n', file);
1799 }
1800 fprintf(file, "%s\n", prefix);
1801 free(data);
1802 }
1803
1804 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1805 {
1806 fprintf(file, "%sGIT binary patch\n", prefix);
1807 emit_binary_diff_body(file, one, two, prefix);
1808 emit_binary_diff_body(file, two, one, prefix);
1809 }
1810
1811 static void diff_filespec_load_driver(struct diff_filespec *one)
1812 {
1813 /* Use already-loaded driver */
1814 if (one->driver)
1815 return;
1816
1817 if (S_ISREG(one->mode))
1818 one->driver = userdiff_find_by_path(one->path);
1819
1820 /* Fallback to default settings */
1821 if (!one->driver)
1822 one->driver = userdiff_find_by_name("default");
1823 }
1824
1825 int diff_filespec_is_binary(struct diff_filespec *one)
1826 {
1827 if (one->is_binary == -1) {
1828 diff_filespec_load_driver(one);
1829 if (one->driver->binary != -1)
1830 one->is_binary = one->driver->binary;
1831 else {
1832 if (!one->data && DIFF_FILE_VALID(one))
1833 diff_populate_filespec(one, 0);
1834 if (one->data)
1835 one->is_binary = buffer_is_binary(one->data,
1836 one->size);
1837 if (one->is_binary == -1)
1838 one->is_binary = 0;
1839 }
1840 }
1841 return one->is_binary;
1842 }
1843
1844 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1845 {
1846 diff_filespec_load_driver(one);
1847 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1848 }
1849
1850 static const char *userdiff_word_regex(struct diff_filespec *one)
1851 {
1852 diff_filespec_load_driver(one);
1853 return one->driver->word_regex;
1854 }
1855
1856 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1857 {
1858 if (!options->a_prefix)
1859 options->a_prefix = a;
1860 if (!options->b_prefix)
1861 options->b_prefix = b;
1862 }
1863
1864 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1865 {
1866 if (!DIFF_FILE_VALID(one))
1867 return NULL;
1868
1869 diff_filespec_load_driver(one);
1870 if (!one->driver->textconv)
1871 return NULL;
1872
1873 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1874 struct notes_cache *c = xmalloc(sizeof(*c));
1875 struct strbuf name = STRBUF_INIT;
1876
1877 strbuf_addf(&name, "textconv/%s", one->driver->name);
1878 notes_cache_init(c, name.buf, one->driver->textconv);
1879 one->driver->textconv_cache = c;
1880 }
1881
1882 return one->driver;
1883 }
1884
1885 static void builtin_diff(const char *name_a,
1886 const char *name_b,
1887 struct diff_filespec *one,
1888 struct diff_filespec *two,
1889 const char *xfrm_msg,
1890 int must_show_header,
1891 struct diff_options *o,
1892 int complete_rewrite)
1893 {
1894 mmfile_t mf1, mf2;
1895 const char *lbl[2];
1896 char *a_one, *b_two;
1897 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1898 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1899 const char *a_prefix, *b_prefix;
1900 struct userdiff_driver *textconv_one = NULL;
1901 struct userdiff_driver *textconv_two = NULL;
1902 struct strbuf header = STRBUF_INIT;
1903 struct strbuf *msgbuf;
1904 char *line_prefix = "";
1905
1906 if (o->output_prefix) {
1907 msgbuf = o->output_prefix(o, o->output_prefix_data);
1908 line_prefix = msgbuf->buf;
1909 }
1910
1911 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1912 (!one->mode || S_ISGITLINK(one->mode)) &&
1913 (!two->mode || S_ISGITLINK(two->mode))) {
1914 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1915 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1916 show_submodule_summary(o->file, one ? one->path : two->path,
1917 one->sha1, two->sha1, two->dirty_submodule,
1918 del, add, reset);
1919 return;
1920 }
1921
1922 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1923 textconv_one = get_textconv(one);
1924 textconv_two = get_textconv(two);
1925 }
1926
1927 diff_set_mnemonic_prefix(o, "a/", "b/");
1928 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1929 a_prefix = o->b_prefix;
1930 b_prefix = o->a_prefix;
1931 } else {
1932 a_prefix = o->a_prefix;
1933 b_prefix = o->b_prefix;
1934 }
1935
1936 /* Never use a non-valid filename anywhere if at all possible */
1937 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1938 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1939
1940 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1941 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1942 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1943 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1944 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1945 if (lbl[0][0] == '/') {
1946 /* /dev/null */
1947 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1948 if (xfrm_msg)
1949 strbuf_addstr(&header, xfrm_msg);
1950 must_show_header = 1;
1951 }
1952 else if (lbl[1][0] == '/') {
1953 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1954 if (xfrm_msg)
1955 strbuf_addstr(&header, xfrm_msg);
1956 must_show_header = 1;
1957 }
1958 else {
1959 if (one->mode != two->mode) {
1960 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1961 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1962 must_show_header = 1;
1963 }
1964 if (xfrm_msg)
1965 strbuf_addstr(&header, xfrm_msg);
1966
1967 /*
1968 * we do not run diff between different kind
1969 * of objects.
1970 */
1971 if ((one->mode ^ two->mode) & S_IFMT)
1972 goto free_ab_and_return;
1973 if (complete_rewrite &&
1974 (textconv_one || !diff_filespec_is_binary(one)) &&
1975 (textconv_two || !diff_filespec_is_binary(two))) {
1976 fprintf(o->file, "%s", header.buf);
1977 strbuf_reset(&header);
1978 emit_rewrite_diff(name_a, name_b, one, two,
1979 textconv_one, textconv_two, o);
1980 o->found_changes = 1;
1981 goto free_ab_and_return;
1982 }
1983 }
1984
1985 if (!DIFF_OPT_TST(o, TEXT) &&
1986 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1987 (!textconv_two && diff_filespec_is_binary(two)) )) {
1988 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1989 die("unable to read files to diff");
1990 /* Quite common confusing case */
1991 if (mf1.size == mf2.size &&
1992 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
1993 if (must_show_header)
1994 fprintf(o->file, "%s", header.buf);
1995 goto free_ab_and_return;
1996 }
1997 fprintf(o->file, "%s", header.buf);
1998 strbuf_reset(&header);
1999 if (DIFF_OPT_TST(o, BINARY))
2000 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2001 else
2002 fprintf(o->file, "%sBinary files %s and %s differ\n",
2003 line_prefix, lbl[0], lbl[1]);
2004 o->found_changes = 1;
2005 }
2006 else {
2007 /* Crazy xdl interfaces.. */
2008 const char *diffopts = getenv("GIT_DIFF_OPTS");
2009 xpparam_t xpp;
2010 xdemitconf_t xecfg;
2011 struct emit_callback ecbdata;
2012 const struct userdiff_funcname *pe;
2013
2014 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2015 fprintf(o->file, "%s", header.buf);
2016 strbuf_reset(&header);
2017 }
2018
2019 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2020 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2021
2022 pe = diff_funcname_pattern(one);
2023 if (!pe)
2024 pe = diff_funcname_pattern(two);
2025
2026 memset(&xpp, 0, sizeof(xpp));
2027 memset(&xecfg, 0, sizeof(xecfg));
2028 memset(&ecbdata, 0, sizeof(ecbdata));
2029 ecbdata.label_path = lbl;
2030 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2031 ecbdata.found_changesp = &o->found_changes;
2032 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2033 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2034 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2035 ecbdata.opt = o;
2036 ecbdata.header = header.len ? &header : NULL;
2037 xpp.flags = o->xdl_opts;
2038 xecfg.ctxlen = o->context;
2039 xecfg.interhunkctxlen = o->interhunkcontext;
2040 xecfg.flags = XDL_EMIT_FUNCNAMES;
2041 if (pe)
2042 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2043 if (!diffopts)
2044 ;
2045 else if (!prefixcmp(diffopts, "--unified="))
2046 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2047 else if (!prefixcmp(diffopts, "-u"))
2048 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2049 if (o->word_diff) {
2050 int i;
2051
2052 ecbdata.diff_words =
2053 xcalloc(1, sizeof(struct diff_words_data));
2054 ecbdata.diff_words->type = o->word_diff;
2055 ecbdata.diff_words->opt = o;
2056 if (!o->word_regex)
2057 o->word_regex = userdiff_word_regex(one);
2058 if (!o->word_regex)
2059 o->word_regex = userdiff_word_regex(two);
2060 if (!o->word_regex)
2061 o->word_regex = diff_word_regex_cfg;
2062 if (o->word_regex) {
2063 ecbdata.diff_words->word_regex = (regex_t *)
2064 xmalloc(sizeof(regex_t));
2065 if (regcomp(ecbdata.diff_words->word_regex,
2066 o->word_regex,
2067 REG_EXTENDED | REG_NEWLINE))
2068 die ("Invalid regular expression: %s",
2069 o->word_regex);
2070 }
2071 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2072 if (o->word_diff == diff_words_styles[i].type) {
2073 ecbdata.diff_words->style =
2074 &diff_words_styles[i];
2075 break;
2076 }
2077 }
2078 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2079 struct diff_words_style *st = ecbdata.diff_words->style;
2080 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2081 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2082 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2083 }
2084 }
2085 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2086 &xpp, &xecfg);
2087 if (o->word_diff)
2088 free_diff_words_data(&ecbdata);
2089 if (textconv_one)
2090 free(mf1.ptr);
2091 if (textconv_two)
2092 free(mf2.ptr);
2093 xdiff_clear_find_func(&xecfg);
2094 }
2095
2096 free_ab_and_return:
2097 strbuf_release(&header);
2098 diff_free_filespec_data(one);
2099 diff_free_filespec_data(two);
2100 free(a_one);
2101 free(b_two);
2102 return;
2103 }
2104
2105 static void builtin_diffstat(const char *name_a, const char *name_b,
2106 struct diff_filespec *one,
2107 struct diff_filespec *two,
2108 struct diffstat_t *diffstat,
2109 struct diff_options *o,
2110 int complete_rewrite)
2111 {
2112 mmfile_t mf1, mf2;
2113 struct diffstat_file *data;
2114
2115 data = diffstat_add(diffstat, name_a, name_b);
2116
2117 if (!one || !two) {
2118 data->is_unmerged = 1;
2119 return;
2120 }
2121
2122 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2123 data->is_binary = 1;
2124 data->added = diff_filespec_size(two);
2125 data->deleted = diff_filespec_size(one);
2126 }
2127
2128 else if (complete_rewrite) {
2129 diff_populate_filespec(one, 0);
2130 diff_populate_filespec(two, 0);
2131 data->deleted = count_lines(one->data, one->size);
2132 data->added = count_lines(two->data, two->size);
2133 }
2134
2135 else {
2136 /* Crazy xdl interfaces.. */
2137 xpparam_t xpp;
2138 xdemitconf_t xecfg;
2139
2140 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2141 die("unable to read files to diff");
2142
2143 memset(&xpp, 0, sizeof(xpp));
2144 memset(&xecfg, 0, sizeof(xecfg));
2145 xpp.flags = o->xdl_opts;
2146 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2147 &xpp, &xecfg);
2148 }
2149
2150 diff_free_filespec_data(one);
2151 diff_free_filespec_data(two);
2152 }
2153
2154 static void builtin_checkdiff(const char *name_a, const char *name_b,
2155 const char *attr_path,
2156 struct diff_filespec *one,
2157 struct diff_filespec *two,
2158 struct diff_options *o)
2159 {
2160 mmfile_t mf1, mf2;
2161 struct checkdiff_t data;
2162
2163 if (!two)
2164 return;
2165
2166 memset(&data, 0, sizeof(data));
2167 data.filename = name_b ? name_b : name_a;
2168 data.lineno = 0;
2169 data.o = o;
2170 data.ws_rule = whitespace_rule(attr_path);
2171 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2172
2173 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2174 die("unable to read files to diff");
2175
2176 /*
2177 * All the other codepaths check both sides, but not checking
2178 * the "old" side here is deliberate. We are checking the newly
2179 * introduced changes, and as long as the "new" side is text, we
2180 * can and should check what it introduces.
2181 */
2182 if (diff_filespec_is_binary(two))
2183 goto free_and_return;
2184 else {
2185 /* Crazy xdl interfaces.. */
2186 xpparam_t xpp;
2187 xdemitconf_t xecfg;
2188
2189 memset(&xpp, 0, sizeof(xpp));
2190 memset(&xecfg, 0, sizeof(xecfg));
2191 xecfg.ctxlen = 1; /* at least one context line */
2192 xpp.flags = 0;
2193 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2194 &xpp, &xecfg);
2195
2196 if (data.ws_rule & WS_BLANK_AT_EOF) {
2197 struct emit_callback ecbdata;
2198 int blank_at_eof;
2199
2200 ecbdata.ws_rule = data.ws_rule;
2201 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2202 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2203
2204 if (blank_at_eof) {
2205 static char *err;
2206 if (!err)
2207 err = whitespace_error_string(WS_BLANK_AT_EOF);
2208 fprintf(o->file, "%s:%d: %s.\n",
2209 data.filename, blank_at_eof, err);
2210 data.status = 1; /* report errors */
2211 }
2212 }
2213 }
2214 free_and_return:
2215 diff_free_filespec_data(one);
2216 diff_free_filespec_data(two);
2217 if (data.status)
2218 DIFF_OPT_SET(o, CHECK_FAILED);
2219 }
2220
2221 struct diff_filespec *alloc_filespec(const char *path)
2222 {
2223 int namelen = strlen(path);
2224 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2225
2226 memset(spec, 0, sizeof(*spec));
2227 spec->path = (char *)(spec + 1);
2228 memcpy(spec->path, path, namelen+1);
2229 spec->count = 1;
2230 spec->is_binary = -1;
2231 return spec;
2232 }
2233
2234 void free_filespec(struct diff_filespec *spec)
2235 {
2236 if (!--spec->count) {
2237 diff_free_filespec_data(spec);
2238 free(spec);
2239 }
2240 }
2241
2242 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2243 unsigned short mode)
2244 {
2245 if (mode) {
2246 spec->mode = canon_mode(mode);
2247 hashcpy(spec->sha1, sha1);
2248 spec->sha1_valid = !is_null_sha1(sha1);
2249 }
2250 }
2251
2252 /*
2253 * Given a name and sha1 pair, if the index tells us the file in
2254 * the work tree has that object contents, return true, so that
2255 * prepare_temp_file() does not have to inflate and extract.
2256 */
2257 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2258 {
2259 struct cache_entry *ce;
2260 struct stat st;
2261 int pos, len;
2262
2263 /*
2264 * We do not read the cache ourselves here, because the
2265 * benchmark with my previous version that always reads cache
2266 * shows that it makes things worse for diff-tree comparing
2267 * two linux-2.6 kernel trees in an already checked out work
2268 * tree. This is because most diff-tree comparisons deal with
2269 * only a small number of files, while reading the cache is
2270 * expensive for a large project, and its cost outweighs the
2271 * savings we get by not inflating the object to a temporary
2272 * file. Practically, this code only helps when we are used
2273 * by diff-cache --cached, which does read the cache before
2274 * calling us.
2275 */
2276 if (!active_cache)
2277 return 0;
2278
2279 /* We want to avoid the working directory if our caller
2280 * doesn't need the data in a normal file, this system
2281 * is rather slow with its stat/open/mmap/close syscalls,
2282 * and the object is contained in a pack file. The pack
2283 * is probably already open and will be faster to obtain
2284 * the data through than the working directory. Loose
2285 * objects however would tend to be slower as they need
2286 * to be individually opened and inflated.
2287 */
2288 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2289 return 0;
2290
2291 len = strlen(name);
2292 pos = cache_name_pos(name, len);
2293 if (pos < 0)
2294 return 0;
2295 ce = active_cache[pos];
2296
2297 /*
2298 * This is not the sha1 we are looking for, or
2299 * unreusable because it is not a regular file.
2300 */
2301 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2302 return 0;
2303
2304 /*
2305 * If ce is marked as "assume unchanged", there is no
2306 * guarantee that work tree matches what we are looking for.
2307 */
2308 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2309 return 0;
2310
2311 /*
2312 * If ce matches the file in the work tree, we can reuse it.
2313 */
2314 if (ce_uptodate(ce) ||
2315 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2316 return 1;
2317
2318 return 0;
2319 }
2320
2321 static int populate_from_stdin(struct diff_filespec *s)
2322 {
2323 struct strbuf buf = STRBUF_INIT;
2324 size_t size = 0;
2325
2326 if (strbuf_read(&buf, 0, 0) < 0)
2327 return error("error while reading from stdin %s",
2328 strerror(errno));
2329
2330 s->should_munmap = 0;
2331 s->data = strbuf_detach(&buf, &size);
2332 s->size = size;
2333 s->should_free = 1;
2334 return 0;
2335 }
2336
2337 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2338 {
2339 int len;
2340 char *data = xmalloc(100), *dirty = "";
2341
2342 /* Are we looking at the work tree? */
2343 if (s->dirty_submodule)
2344 dirty = "-dirty";
2345
2346 len = snprintf(data, 100,
2347 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2348 s->data = data;
2349 s->size = len;
2350 s->should_free = 1;
2351 if (size_only) {
2352 s->data = NULL;
2353 free(data);
2354 }
2355 return 0;
2356 }
2357
2358 /*
2359 * While doing rename detection and pickaxe operation, we may need to
2360 * grab the data for the blob (or file) for our own in-core comparison.
2361 * diff_filespec has data and size fields for this purpose.
2362 */
2363 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2364 {
2365 int err = 0;
2366 if (!DIFF_FILE_VALID(s))
2367 die("internal error: asking to populate invalid file.");
2368 if (S_ISDIR(s->mode))
2369 return -1;
2370
2371 if (s->data)
2372 return 0;
2373
2374 if (size_only && 0 < s->size)
2375 return 0;
2376
2377 if (S_ISGITLINK(s->mode))
2378 return diff_populate_gitlink(s, size_only);
2379
2380 if (!s->sha1_valid ||
2381 reuse_worktree_file(s->path, s->sha1, 0)) {
2382 struct strbuf buf = STRBUF_INIT;
2383 struct stat st;
2384 int fd;
2385
2386 if (!strcmp(s->path, "-"))
2387 return populate_from_stdin(s);
2388
2389 if (lstat(s->path, &st) < 0) {
2390 if (errno == ENOENT) {
2391 err_empty:
2392 err = -1;
2393 empty:
2394 s->data = (char *)"";
2395 s->size = 0;
2396 return err;
2397 }
2398 }
2399 s->size = xsize_t(st.st_size);
2400 if (!s->size)
2401 goto empty;
2402 if (S_ISLNK(st.st_mode)) {
2403 struct strbuf sb = STRBUF_INIT;
2404
2405 if (strbuf_readlink(&sb, s->path, s->size))
2406 goto err_empty;
2407 s->size = sb.len;
2408 s->data = strbuf_detach(&sb, NULL);
2409 s->should_free = 1;
2410 return 0;
2411 }
2412 if (size_only)
2413 return 0;
2414 fd = open(s->path, O_RDONLY);
2415 if (fd < 0)
2416 goto err_empty;
2417 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2418 close(fd);
2419 s->should_munmap = 1;
2420
2421 /*
2422 * Convert from working tree format to canonical git format
2423 */
2424 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2425 size_t size = 0;
2426 munmap(s->data, s->size);
2427 s->should_munmap = 0;
2428 s->data = strbuf_detach(&buf, &size);
2429 s->size = size;
2430 s->should_free = 1;
2431 }
2432 }
2433 else {
2434 enum object_type type;
2435 if (size_only) {
2436 type = sha1_object_info(s->sha1, &s->size);
2437 if (type < 0)
2438 die("unable to read %s", sha1_to_hex(s->sha1));
2439 } else {
2440 s->data = read_sha1_file(s->sha1, &type, &s->size);
2441 if (!s->data)
2442 die("unable to read %s", sha1_to_hex(s->sha1));
2443 s->should_free = 1;
2444 }
2445 }
2446 return 0;
2447 }
2448
2449 void diff_free_filespec_blob(struct diff_filespec *s)
2450 {
2451 if (s->should_free)
2452 free(s->data);
2453 else if (s->should_munmap)
2454 munmap(s->data, s->size);
2455
2456 if (s->should_free || s->should_munmap) {
2457 s->should_free = s->should_munmap = 0;
2458 s->data = NULL;
2459 }
2460 }
2461
2462 void diff_free_filespec_data(struct diff_filespec *s)
2463 {
2464 diff_free_filespec_blob(s);
2465 free(s->cnt_data);
2466 s->cnt_data = NULL;
2467 }
2468
2469 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2470 void *blob,
2471 unsigned long size,
2472 const unsigned char *sha1,
2473 int mode)
2474 {
2475 int fd;
2476 struct strbuf buf = STRBUF_INIT;
2477 struct strbuf template = STRBUF_INIT;
2478 char *path_dup = xstrdup(path);
2479 const char *base = basename(path_dup);
2480
2481 /* Generate "XXXXXX_basename.ext" */
2482 strbuf_addstr(&template, "XXXXXX_");
2483 strbuf_addstr(&template, base);
2484
2485 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2486 strlen(base) + 1);
2487 if (fd < 0)
2488 die_errno("unable to create temp-file");
2489 if (convert_to_working_tree(path,
2490 (const char *)blob, (size_t)size, &buf)) {
2491 blob = buf.buf;
2492 size = buf.len;
2493 }
2494 if (write_in_full(fd, blob, size) != size)
2495 die_errno("unable to write temp-file");
2496 close(fd);
2497 temp->name = temp->tmp_path;
2498 strcpy(temp->hex, sha1_to_hex(sha1));
2499 temp->hex[40] = 0;
2500 sprintf(temp->mode, "%06o", mode);
2501 strbuf_release(&buf);
2502 strbuf_release(&template);
2503 free(path_dup);
2504 }
2505
2506 static struct diff_tempfile *prepare_temp_file(const char *name,
2507 struct diff_filespec *one)
2508 {
2509 struct diff_tempfile *temp = claim_diff_tempfile();
2510
2511 if (!DIFF_FILE_VALID(one)) {
2512 not_a_valid_file:
2513 /* A '-' entry produces this for file-2, and
2514 * a '+' entry produces this for file-1.
2515 */
2516 temp->name = "/dev/null";
2517 strcpy(temp->hex, ".");
2518 strcpy(temp->mode, ".");
2519 return temp;
2520 }
2521
2522 if (!remove_tempfile_installed) {
2523 atexit(remove_tempfile);
2524 sigchain_push_common(remove_tempfile_on_signal);
2525 remove_tempfile_installed = 1;
2526 }
2527
2528 if (!one->sha1_valid ||
2529 reuse_worktree_file(name, one->sha1, 1)) {
2530 struct stat st;
2531 if (lstat(name, &st) < 0) {
2532 if (errno == ENOENT)
2533 goto not_a_valid_file;
2534 die_errno("stat(%s)", name);
2535 }
2536 if (S_ISLNK(st.st_mode)) {
2537 struct strbuf sb = STRBUF_INIT;
2538 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2539 die_errno("readlink(%s)", name);
2540 prep_temp_blob(name, temp, sb.buf, sb.len,
2541 (one->sha1_valid ?
2542 one->sha1 : null_sha1),
2543 (one->sha1_valid ?
2544 one->mode : S_IFLNK));
2545 strbuf_release(&sb);
2546 }
2547 else {
2548 /* we can borrow from the file in the work tree */
2549 temp->name = name;
2550 if (!one->sha1_valid)
2551 strcpy(temp->hex, sha1_to_hex(null_sha1));
2552 else
2553 strcpy(temp->hex, sha1_to_hex(one->sha1));
2554 /* Even though we may sometimes borrow the
2555 * contents from the work tree, we always want
2556 * one->mode. mode is trustworthy even when
2557 * !(one->sha1_valid), as long as
2558 * DIFF_FILE_VALID(one).
2559 */
2560 sprintf(temp->mode, "%06o", one->mode);
2561 }
2562 return temp;
2563 }
2564 else {
2565 if (diff_populate_filespec(one, 0))
2566 die("cannot read data blob for %s", one->path);
2567 prep_temp_blob(name, temp, one->data, one->size,
2568 one->sha1, one->mode);
2569 }
2570 return temp;
2571 }
2572
2573 /* An external diff command takes:
2574 *
2575 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2576 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2577 *
2578 */
2579 static void run_external_diff(const char *pgm,
2580 const char *name,
2581 const char *other,
2582 struct diff_filespec *one,
2583 struct diff_filespec *two,
2584 const char *xfrm_msg,
2585 int complete_rewrite)
2586 {
2587 const char *spawn_arg[10];
2588 int retval;
2589 const char **arg = &spawn_arg[0];
2590
2591 if (one && two) {
2592 struct diff_tempfile *temp_one, *temp_two;
2593 const char *othername = (other ? other : name);
2594 temp_one = prepare_temp_file(name, one);
2595 temp_two = prepare_temp_file(othername, two);
2596 *arg++ = pgm;
2597 *arg++ = name;
2598 *arg++ = temp_one->name;
2599 *arg++ = temp_one->hex;
2600 *arg++ = temp_one->mode;
2601 *arg++ = temp_two->name;
2602 *arg++ = temp_two->hex;
2603 *arg++ = temp_two->mode;
2604 if (other) {
2605 *arg++ = other;
2606 *arg++ = xfrm_msg;
2607 }
2608 } else {
2609 *arg++ = pgm;
2610 *arg++ = name;
2611 }
2612 *arg = NULL;
2613 fflush(NULL);
2614 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2615 remove_tempfile();
2616 if (retval) {
2617 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2618 exit(1);
2619 }
2620 }
2621
2622 static int similarity_index(struct diff_filepair *p)
2623 {
2624 return p->score * 100 / MAX_SCORE;
2625 }
2626
2627 static void fill_metainfo(struct strbuf *msg,
2628 const char *name,
2629 const char *other,
2630 struct diff_filespec *one,
2631 struct diff_filespec *two,
2632 struct diff_options *o,
2633 struct diff_filepair *p,
2634 int *must_show_header,
2635 int use_color)
2636 {
2637 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2638 const char *reset = diff_get_color(use_color, DIFF_RESET);
2639 struct strbuf *msgbuf;
2640 char *line_prefix = "";
2641
2642 *must_show_header = 1;
2643 if (o->output_prefix) {
2644 msgbuf = o->output_prefix(o, o->output_prefix_data);
2645 line_prefix = msgbuf->buf;
2646 }
2647 strbuf_init(msg, PATH_MAX * 2 + 300);
2648 switch (p->status) {
2649 case DIFF_STATUS_COPIED:
2650 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2651 line_prefix, set, similarity_index(p));
2652 strbuf_addf(msg, "%s\n%s%scopy from ",
2653 reset, line_prefix, set);
2654 quote_c_style(name, msg, NULL, 0);
2655 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2656 quote_c_style(other, msg, NULL, 0);
2657 strbuf_addf(msg, "%s\n", reset);
2658 break;
2659 case DIFF_STATUS_RENAMED:
2660 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2661 line_prefix, set, similarity_index(p));
2662 strbuf_addf(msg, "%s\n%s%srename from ",
2663 reset, line_prefix, set);
2664 quote_c_style(name, msg, NULL, 0);
2665 strbuf_addf(msg, "%s\n%s%srename to ",
2666 reset, line_prefix, set);
2667 quote_c_style(other, msg, NULL, 0);
2668 strbuf_addf(msg, "%s\n", reset);
2669 break;
2670 case DIFF_STATUS_MODIFIED:
2671 if (p->score) {
2672 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2673 line_prefix,
2674 set, similarity_index(p), reset);
2675 break;
2676 }
2677 /* fallthru */
2678 default:
2679 *must_show_header = 0;
2680 }
2681 if (one && two && hashcmp(one->sha1, two->sha1)) {
2682 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2683
2684 if (DIFF_OPT_TST(o, BINARY)) {
2685 mmfile_t mf;
2686 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2687 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2688 abbrev = 40;
2689 }
2690 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2691 find_unique_abbrev(one->sha1, abbrev));
2692 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2693 if (one->mode == two->mode)
2694 strbuf_addf(msg, " %06o", one->mode);
2695 strbuf_addf(msg, "%s\n", reset);
2696 }
2697 }
2698
2699 static void run_diff_cmd(const char *pgm,
2700 const char *name,
2701 const char *other,
2702 const char *attr_path,
2703 struct diff_filespec *one,
2704 struct diff_filespec *two,
2705 struct strbuf *msg,
2706 struct diff_options *o,
2707 struct diff_filepair *p)
2708 {
2709 const char *xfrm_msg = NULL;
2710 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2711 int must_show_header = 0;
2712
2713 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2714 pgm = NULL;
2715 else {
2716 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2717 if (drv && drv->external)
2718 pgm = drv->external;
2719 }
2720
2721 if (msg) {
2722 /*
2723 * don't use colors when the header is intended for an
2724 * external diff driver
2725 */
2726 fill_metainfo(msg, name, other, one, two, o, p,
2727 &must_show_header,
2728 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2729 xfrm_msg = msg->len ? msg->buf : NULL;
2730 }
2731
2732 if (pgm) {
2733 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2734 complete_rewrite);
2735 return;
2736 }
2737 if (one && two)
2738 builtin_diff(name, other ? other : name,
2739 one, two, xfrm_msg, must_show_header,
2740 o, complete_rewrite);
2741 else
2742 fprintf(o->file, "* Unmerged path %s\n", name);
2743 }
2744
2745 static void diff_fill_sha1_info(struct diff_filespec *one)
2746 {
2747 if (DIFF_FILE_VALID(one)) {
2748 if (!one->sha1_valid) {
2749 struct stat st;
2750 if (!strcmp(one->path, "-")) {
2751 hashcpy(one->sha1, null_sha1);
2752 return;
2753 }
2754 if (lstat(one->path, &st) < 0)
2755 die_errno("stat '%s'", one->path);
2756 if (index_path(one->sha1, one->path, &st, 0))
2757 die("cannot hash %s", one->path);
2758 }
2759 }
2760 else
2761 hashclr(one->sha1);
2762 }
2763
2764 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2765 {
2766 /* Strip the prefix but do not molest /dev/null and absolute paths */
2767 if (*namep && **namep != '/') {
2768 *namep += prefix_length;
2769 if (**namep == '/')
2770 ++*namep;
2771 }
2772 if (*otherp && **otherp != '/') {
2773 *otherp += prefix_length;
2774 if (**otherp == '/')
2775 ++*otherp;
2776 }
2777 }
2778
2779 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2780 {
2781 const char *pgm = external_diff();
2782 struct strbuf msg;
2783 struct diff_filespec *one = p->one;
2784 struct diff_filespec *two = p->two;
2785 const char *name;
2786 const char *other;
2787 const char *attr_path;
2788
2789 name = p->one->path;
2790 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2791 attr_path = name;
2792 if (o->prefix_length)
2793 strip_prefix(o->prefix_length, &name, &other);
2794
2795 if (DIFF_PAIR_UNMERGED(p)) {
2796 run_diff_cmd(pgm, name, NULL, attr_path,
2797 NULL, NULL, NULL, o, p);
2798 return;
2799 }
2800
2801 diff_fill_sha1_info(one);
2802 diff_fill_sha1_info(two);
2803
2804 if (!pgm &&
2805 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2806 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2807 /*
2808 * a filepair that changes between file and symlink
2809 * needs to be split into deletion and creation.
2810 */
2811 struct diff_filespec *null = alloc_filespec(two->path);
2812 run_diff_cmd(NULL, name, other, attr_path,
2813 one, null, &msg, o, p);
2814 free(null);
2815 strbuf_release(&msg);
2816
2817 null = alloc_filespec(one->path);
2818 run_diff_cmd(NULL, name, other, attr_path,
2819 null, two, &msg, o, p);
2820 free(null);
2821 }
2822 else
2823 run_diff_cmd(pgm, name, other, attr_path,
2824 one, two, &msg, o, p);
2825
2826 strbuf_release(&msg);
2827 }
2828
2829 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2830 struct diffstat_t *diffstat)
2831 {
2832 const char *name;
2833 const char *other;
2834 int complete_rewrite = 0;
2835
2836 if (DIFF_PAIR_UNMERGED(p)) {
2837 /* unmerged */
2838 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2839 return;
2840 }
2841
2842 name = p->one->path;
2843 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2844
2845 if (o->prefix_length)
2846 strip_prefix(o->prefix_length, &name, &other);
2847
2848 diff_fill_sha1_info(p->one);
2849 diff_fill_sha1_info(p->two);
2850
2851 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2852 complete_rewrite = 1;
2853 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2854 }
2855
2856 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2857 {
2858 const char *name;
2859 const char *other;
2860 const char *attr_path;
2861
2862 if (DIFF_PAIR_UNMERGED(p)) {
2863 /* unmerged */
2864 return;
2865 }
2866
2867 name = p->one->path;
2868 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2869 attr_path = other ? other : name;
2870
2871 if (o->prefix_length)
2872 strip_prefix(o->prefix_length, &name, &other);
2873
2874 diff_fill_sha1_info(p->one);
2875 diff_fill_sha1_info(p->two);
2876
2877 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2878 }
2879
2880 void diff_setup(struct diff_options *options)
2881 {
2882 memcpy(options, &default_diff_options, sizeof(*options));
2883
2884 options->file = stdout;
2885
2886 options->line_termination = '\n';
2887 options->break_opt = -1;
2888 options->rename_limit = -1;
2889 options->dirstat_percent = 3;
2890 options->context = 3;
2891
2892 options->change = diff_change;
2893 options->add_remove = diff_addremove;
2894 if (diff_use_color_default > 0)
2895 DIFF_OPT_SET(options, COLOR_DIFF);
2896 options->detect_rename = diff_detect_rename_default;
2897
2898 if (diff_no_prefix) {
2899 options->a_prefix = options->b_prefix = "";
2900 } else if (!diff_mnemonic_prefix) {
2901 options->a_prefix = "a/";
2902 options->b_prefix = "b/";
2903 }
2904 }
2905
2906 int diff_setup_done(struct diff_options *options)
2907 {
2908 int count = 0;
2909
2910 if (options->output_format & DIFF_FORMAT_NAME)
2911 count++;
2912 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2913 count++;
2914 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2915 count++;
2916 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2917 count++;
2918 if (count > 1)
2919 die("--name-only, --name-status, --check and -s are mutually exclusive");
2920
2921 /*
2922 * Most of the time we can say "there are changes"
2923 * only by checking if there are changed paths, but
2924 * --ignore-whitespace* options force us to look
2925 * inside contents.
2926 */
2927
2928 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2929 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2930 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2931 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2932 else
2933 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2934
2935 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2936 options->detect_rename = DIFF_DETECT_COPY;
2937
2938 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2939 options->prefix = NULL;
2940 if (options->prefix)
2941 options->prefix_length = strlen(options->prefix);
2942 else
2943 options->prefix_length = 0;
2944
2945 if (options->output_format & (DIFF_FORMAT_NAME |
2946 DIFF_FORMAT_NAME_STATUS |
2947 DIFF_FORMAT_CHECKDIFF |
2948 DIFF_FORMAT_NO_OUTPUT))
2949 options->output_format &= ~(DIFF_FORMAT_RAW |
2950 DIFF_FORMAT_NUMSTAT |
2951 DIFF_FORMAT_DIFFSTAT |
2952 DIFF_FORMAT_SHORTSTAT |
2953 DIFF_FORMAT_DIRSTAT |
2954 DIFF_FORMAT_SUMMARY |
2955 DIFF_FORMAT_PATCH);
2956
2957 /*
2958 * These cases always need recursive; we do not drop caller-supplied
2959 * recursive bits for other formats here.
2960 */
2961 if (options->output_format & (DIFF_FORMAT_PATCH |
2962 DIFF_FORMAT_NUMSTAT |
2963 DIFF_FORMAT_DIFFSTAT |
2964 DIFF_FORMAT_SHORTSTAT |
2965 DIFF_FORMAT_DIRSTAT |
2966 DIFF_FORMAT_SUMMARY |
2967 DIFF_FORMAT_CHECKDIFF))
2968 DIFF_OPT_SET(options, RECURSIVE);
2969 /*
2970 * Also pickaxe would not work very well if you do not say recursive
2971 */
2972 if (options->pickaxe)
2973 DIFF_OPT_SET(options, RECURSIVE);
2974 /*
2975 * When patches are generated, submodules diffed against the work tree
2976 * must be checked for dirtiness too so it can be shown in the output
2977 */
2978 if (options->output_format & DIFF_FORMAT_PATCH)
2979 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2980
2981 if (options->detect_rename && options->rename_limit < 0)
2982 options->rename_limit = diff_rename_limit_default;
2983 if (options->setup & DIFF_SETUP_USE_CACHE) {
2984 if (!active_cache)
2985 /* read-cache does not die even when it fails
2986 * so it is safe for us to do this here. Also
2987 * it does not smudge active_cache or active_nr
2988 * when it fails, so we do not have to worry about
2989 * cleaning it up ourselves either.
2990 */
2991 read_cache();
2992 }
2993 if (options->abbrev <= 0 || 40 < options->abbrev)
2994 options->abbrev = 40; /* full */
2995
2996 /*
2997 * It does not make sense to show the first hit we happened
2998 * to have found. It does not make sense not to return with
2999 * exit code in such a case either.
3000 */
3001 if (DIFF_OPT_TST(options, QUICK)) {
3002 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3003 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3004 }
3005
3006 return 0;
3007 }
3008
3009 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3010 {
3011 char c, *eq;
3012 int len;
3013
3014 if (*arg != '-')
3015 return 0;
3016 c = *++arg;
3017 if (!c)
3018 return 0;
3019 if (c == arg_short) {
3020 c = *++arg;
3021 if (!c)
3022 return 1;
3023 if (val && isdigit(c)) {
3024 char *end;
3025 int n = strtoul(arg, &end, 10);
3026 if (*end)
3027 return 0;
3028 *val = n;
3029 return 1;
3030 }
3031 return 0;
3032 }
3033 if (c != '-')
3034 return 0;
3035 arg++;
3036 eq = strchr(arg, '=');
3037 if (eq)
3038 len = eq - arg;
3039 else
3040 len = strlen(arg);
3041 if (!len || strncmp(arg, arg_long, len))
3042 return 0;
3043 if (eq) {
3044 int n;
3045 char *end;
3046 if (!isdigit(*++eq))
3047 return 0;
3048 n = strtoul(eq, &end, 10);
3049 if (*end)
3050 return 0;
3051 *val = n;
3052 }
3053 return 1;
3054 }
3055
3056 static int diff_scoreopt_parse(const char *opt);
3057
3058 static inline int short_opt(char opt, const char **argv,
3059 const char **optarg)
3060 {
3061 const char *arg = argv[0];
3062 if (arg[0] != '-' || arg[1] != opt)
3063 return 0;
3064 if (arg[2] != '\0') {
3065 *optarg = arg + 2;
3066 return 1;
3067 }
3068 if (!argv[1])
3069 die("Option '%c' requires a value", opt);
3070 *optarg = argv[1];
3071 return 2;
3072 }
3073
3074 int parse_long_opt(const char *opt, const char **argv,
3075 const char **optarg)
3076 {
3077 const char *arg = argv[0];
3078 if (arg[0] != '-' || arg[1] != '-')
3079 return 0;
3080 arg += strlen("--");
3081 if (prefixcmp(arg, opt))
3082 return 0;
3083 arg += strlen(opt);
3084 if (*arg == '=') { /* sticked form: --option=value */
3085 *optarg = arg + 1;
3086 return 1;
3087 }
3088 if (*arg != '\0')
3089 return 0;
3090 /* separate form: --option value */
3091 if (!argv[1])
3092 die("Option '--%s' requires a value", opt);
3093 *optarg = argv[1];
3094 return 2;
3095 }
3096
3097 static int stat_opt(struct diff_options *options, const char **av)
3098 {
3099 const char *arg = av[0];
3100 char *end;
3101 int width = options->stat_width;
3102 int name_width = options->stat_name_width;
3103 int argcount = 1;
3104
3105 arg += strlen("--stat");
3106 end = (char *)arg;
3107
3108 switch (*arg) {
3109 case '-':
3110 if (!prefixcmp(arg, "-width")) {
3111 arg += strlen("-width");
3112 if (*arg == '=')
3113 width = strtoul(arg + 1, &end, 10);
3114 else if (!*arg && !av[1])
3115 die("Option '--stat-width' requires a value");
3116 else if (!*arg) {
3117 width = strtoul(av[1], &end, 10);
3118 argcount = 2;
3119 }
3120 } else if (!prefixcmp(arg, "-name-width")) {
3121 arg += strlen("-name-width");
3122 if (*arg == '=')
3123 name_width = strtoul(arg + 1, &end, 10);
3124 else if (!*arg && !av[1])
3125 die("Option '--stat-name-width' requires a value");
3126 else if (!*arg) {
3127 name_width = strtoul(av[1], &end, 10);
3128 argcount = 2;
3129 }
3130 }
3131 break;
3132 case '=':
3133 width = strtoul(arg+1, &end, 10);
3134 if (*end == ',')
3135 name_width = strtoul(end+1, &end, 10);
3136 }
3137
3138 /* Important! This checks all the error cases! */
3139 if (*end)
3140 return 0;
3141 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3142 options->stat_name_width = name_width;
3143 options->stat_width = width;
3144 return argcount;
3145 }
3146
3147 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3148 {
3149 const char *arg = av[0];
3150 const char *optarg;
3151 int argcount;
3152
3153 /* Output format options */
3154 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3155 options->output_format |= DIFF_FORMAT_PATCH;
3156 else if (opt_arg(arg, 'U', "unified", &options->context))
3157 options->output_format |= DIFF_FORMAT_PATCH;
3158 else if (!strcmp(arg, "--raw"))
3159 options->output_format |= DIFF_FORMAT_RAW;
3160 else if (!strcmp(arg, "--patch-with-raw"))
3161 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3162 else if (!strcmp(arg, "--numstat"))
3163 options->output_format |= DIFF_FORMAT_NUMSTAT;
3164 else if (!strcmp(arg, "--shortstat"))
3165 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3166 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3167 options->output_format |= DIFF_FORMAT_DIRSTAT;
3168 else if (!strcmp(arg, "--cumulative")) {
3169 options->output_format |= DIFF_FORMAT_DIRSTAT;
3170 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3171 } else if (opt_arg(arg, 0, "dirstat-by-file",
3172 &options->dirstat_percent)) {
3173 options->output_format |= DIFF_FORMAT_DIRSTAT;
3174 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3175 }
3176 else if (!strcmp(arg, "--check"))
3177 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3178 else if (!strcmp(arg, "--summary"))
3179 options->output_format |= DIFF_FORMAT_SUMMARY;
3180 else if (!strcmp(arg, "--patch-with-stat"))
3181 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3182 else if (!strcmp(arg, "--name-only"))
3183 options->output_format |= DIFF_FORMAT_NAME;
3184 else if (!strcmp(arg, "--name-status"))
3185 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3186 else if (!strcmp(arg, "-s"))
3187 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3188 else if (!prefixcmp(arg, "--stat"))
3189 /* --stat, --stat-width, or --stat-name-width */
3190 return stat_opt(options, av);
3191
3192 /* renames options */
3193 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3194 !strcmp(arg, "--break-rewrites")) {
3195 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3196 return error("invalid argument to -B: %s", arg+2);
3197 }
3198 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3199 !strcmp(arg, "--find-renames")) {
3200 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3201 return error("invalid argument to -M: %s", arg+2);
3202 options->detect_rename = DIFF_DETECT_RENAME;
3203 }
3204 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3205 !strcmp(arg, "--find-copies")) {
3206 if (options->detect_rename == DIFF_DETECT_COPY)
3207 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3208 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3209 return error("invalid argument to -C: %s", arg+2);
3210 options->detect_rename = DIFF_DETECT_COPY;
3211 }
3212 else if (!strcmp(arg, "--no-renames"))
3213 options->detect_rename = 0;
3214 else if (!strcmp(arg, "--relative"))
3215 DIFF_OPT_SET(options, RELATIVE_NAME);
3216 else if (!prefixcmp(arg, "--relative=")) {
3217 DIFF_OPT_SET(options, RELATIVE_NAME);
3218 options->prefix = arg + 11;
3219 }
3220
3221 /* xdiff options */
3222 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3223 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3224 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3225 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3226 else if (!strcmp(arg, "--ignore-space-at-eol"))
3227 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3228 else if (!strcmp(arg, "--patience"))
3229 DIFF_XDL_SET(options, PATIENCE_DIFF);
3230
3231 /* flags options */
3232 else if (!strcmp(arg, "--binary")) {
3233 options->output_format |= DIFF_FORMAT_PATCH;
3234 DIFF_OPT_SET(options, BINARY);
3235 }
3236 else if (!strcmp(arg, "--full-index"))
3237 DIFF_OPT_SET(options, FULL_INDEX);
3238 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3239 DIFF_OPT_SET(options, TEXT);
3240 else if (!strcmp(arg, "-R"))
3241 DIFF_OPT_SET(options, REVERSE_DIFF);
3242 else if (!strcmp(arg, "--find-copies-harder"))
3243 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3244 else if (!strcmp(arg, "--follow"))
3245 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3246 else if (!strcmp(arg, "--color"))
3247 DIFF_OPT_SET(options, COLOR_DIFF);
3248 else if (!prefixcmp(arg, "--color=")) {
3249 int value = git_config_colorbool(NULL, arg+8, -1);
3250 if (value == 0)
3251 DIFF_OPT_CLR(options, COLOR_DIFF);
3252 else if (value > 0)
3253 DIFF_OPT_SET(options, COLOR_DIFF);
3254 else
3255 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3256 }
3257 else if (!strcmp(arg, "--no-color"))
3258 DIFF_OPT_CLR(options, COLOR_DIFF);
3259 else if (!strcmp(arg, "--color-words")) {
3260 DIFF_OPT_SET(options, COLOR_DIFF);
3261 options->word_diff = DIFF_WORDS_COLOR;
3262 }
3263 else if (!prefixcmp(arg, "--color-words=")) {
3264 DIFF_OPT_SET(options, COLOR_DIFF);
3265 options->word_diff = DIFF_WORDS_COLOR;
3266 options->word_regex = arg + 14;
3267 }
3268 else if (!strcmp(arg, "--word-diff")) {
3269 if (options->word_diff == DIFF_WORDS_NONE)
3270 options->word_diff = DIFF_WORDS_PLAIN;
3271 }
3272 else if (!prefixcmp(arg, "--word-diff=")) {
3273 const char *type = arg + 12;
3274 if (!strcmp(type, "plain"))
3275 options->word_diff = DIFF_WORDS_PLAIN;
3276 else if (!strcmp(type, "color")) {
3277 DIFF_OPT_SET(options, COLOR_DIFF);
3278 options->word_diff = DIFF_WORDS_COLOR;
3279 }
3280 else if (!strcmp(type, "porcelain"))
3281 options->word_diff = DIFF_WORDS_PORCELAIN;
3282 else if (!strcmp(type, "none"))
3283 options->word_diff = DIFF_WORDS_NONE;
3284 else
3285 die("bad --word-diff argument: %s", type);
3286 }
3287 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3288 if (options->word_diff == DIFF_WORDS_NONE)
3289 options->word_diff = DIFF_WORDS_PLAIN;
3290 options->word_regex = optarg;
3291 return argcount;
3292 }
3293 else if (!strcmp(arg, "--exit-code"))
3294 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3295 else if (!strcmp(arg, "--quiet"))
3296 DIFF_OPT_SET(options, QUICK);
3297 else if (!strcmp(arg, "--ext-diff"))
3298 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3299 else if (!strcmp(arg, "--no-ext-diff"))
3300 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3301 else if (!strcmp(arg, "--textconv"))
3302 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3303 else if (!strcmp(arg, "--no-textconv"))
3304 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3305 else if (!strcmp(arg, "--ignore-submodules")) {
3306 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3307 handle_ignore_submodules_arg(options, "all");
3308 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3309 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3310 handle_ignore_submodules_arg(options, arg + 20);
3311 } else if (!strcmp(arg, "--submodule"))
3312 DIFF_OPT_SET(options, SUBMODULE_LOG);
3313 else if (!prefixcmp(arg, "--submodule=")) {
3314 if (!strcmp(arg + 12, "log"))
3315 DIFF_OPT_SET(options, SUBMODULE_LOG);
3316 }
3317
3318 /* misc options */
3319 else if (!strcmp(arg, "-z"))
3320 options->line_termination = 0;
3321 else if ((argcount = short_opt('l', av, &optarg))) {
3322 options->rename_limit = strtoul(optarg, NULL, 10);
3323 return argcount;
3324 }
3325 else if ((argcount = short_opt('S', av, &optarg))) {
3326 options->pickaxe = optarg;
3327 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3328 return argcount;
3329 } else if ((argcount = short_opt('G', av, &optarg))) {
3330 options->pickaxe = optarg;
3331 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3332 return argcount;
3333 }
3334 else if (!strcmp(arg, "--pickaxe-all"))
3335 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3336 else if (!strcmp(arg, "--pickaxe-regex"))
3337 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3338 else if ((argcount = short_opt('O', av, &optarg))) {
3339 options->orderfile = optarg;
3340 return argcount;
3341 }
3342 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3343 options->filter = optarg;
3344 return argcount;
3345 }
3346 else if (!strcmp(arg, "--abbrev"))
3347 options->abbrev = DEFAULT_ABBREV;
3348 else if (!prefixcmp(arg, "--abbrev=")) {
3349 options->abbrev = strtoul(arg + 9, NULL, 10);
3350 if (options->abbrev < MINIMUM_ABBREV)
3351 options->abbrev = MINIMUM_ABBREV;
3352 else if (40 < options->abbrev)
3353 options->abbrev = 40;
3354 }
3355 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3356 options->a_prefix = optarg;
3357 return argcount;
3358 }
3359 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3360 options->b_prefix = optarg;
3361 return argcount;
3362 }
3363 else if (!strcmp(arg, "--no-prefix"))
3364 options->a_prefix = options->b_prefix = "";
3365 else if (opt_arg(arg, '\0', "inter-hunk-context",
3366 &options->interhunkcontext))
3367 ;
3368 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3369 options->file = fopen(optarg, "w");
3370 if (!options->file)
3371 die_errno("Could not open '%s'", optarg);
3372 options->close_file = 1;
3373 return argcount;
3374 } else
3375 return 0;
3376 return 1;
3377 }
3378
3379 int parse_rename_score(const char **cp_p)
3380 {
3381 unsigned long num, scale;
3382 int ch, dot;
3383 const char *cp = *cp_p;
3384
3385 num = 0;
3386 scale = 1;
3387 dot = 0;
3388 for (;;) {
3389 ch = *cp;
3390 if ( !dot && ch == '.' ) {
3391 scale = 1;
3392 dot = 1;
3393 } else if ( ch == '%' ) {
3394 scale = dot ? scale*100 : 100;
3395 cp++; /* % is always at the end */
3396 break;
3397 } else if ( ch >= '0' && ch <= '9' ) {
3398 if ( scale < 100000 ) {
3399 scale *= 10;
3400 num = (num*10) + (ch-'0');
3401 }
3402 } else {
3403 break;
3404 }
3405 cp++;
3406 }
3407 *cp_p = cp;
3408
3409 /* user says num divided by scale and we say internally that
3410 * is MAX_SCORE * num / scale.
3411 */
3412 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3413 }
3414
3415 static int diff_scoreopt_parse(const char *opt)
3416 {
3417 int opt1, opt2, cmd;
3418
3419 if (*opt++ != '-')
3420 return -1;
3421 cmd = *opt++;
3422 if (cmd == '-') {
3423 /* convert the long-form arguments into short-form versions */
3424 if (!prefixcmp(opt, "break-rewrites")) {
3425 opt += strlen("break-rewrites");
3426 if (*opt == 0 || *opt++ == '=')
3427 cmd = 'B';
3428 } else if (!prefixcmp(opt, "find-copies")) {
3429 opt += strlen("find-copies");
3430 if (*opt == 0 || *opt++ == '=')
3431 cmd = 'C';
3432 } else if (!prefixcmp(opt, "find-renames")) {
3433 opt += strlen("find-renames");
3434 if (*opt == 0 || *opt++ == '=')
3435 cmd = 'M';
3436 }
3437 }
3438 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3439 return -1; /* that is not a -M, -C nor -B option */
3440
3441 opt1 = parse_rename_score(&opt);
3442 if (cmd != 'B')
3443 opt2 = 0;
3444 else {
3445 if (*opt == 0)
3446 opt2 = 0;
3447 else if (*opt != '/')
3448 return -1; /* we expect -B80/99 or -B80 */
3449 else {
3450 opt++;
3451 opt2 = parse_rename_score(&opt);
3452 }
3453 }
3454 if (*opt != 0)
3455 return -1;
3456 return opt1 | (opt2 << 16);
3457 }
3458
3459 struct diff_queue_struct diff_queued_diff;
3460
3461 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3462 {
3463 if (queue->alloc <= queue->nr) {
3464 queue->alloc = alloc_nr(queue->alloc);
3465 queue->queue = xrealloc(queue->queue,
3466 sizeof(dp) * queue->alloc);
3467 }
3468 queue->queue[queue->nr++] = dp;
3469 }
3470
3471 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3472 struct diff_filespec *one,
3473 struct diff_filespec *two)
3474 {
3475 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3476 dp->one = one;
3477 dp->two = two;
3478 if (queue)
3479 diff_q(queue, dp);
3480 return dp;
3481 }
3482
3483 void diff_free_filepair(struct diff_filepair *p)
3484 {
3485 free_filespec(p->one);
3486 free_filespec(p->two);
3487 free(p);
3488 }
3489
3490 /* This is different from find_unique_abbrev() in that
3491 * it stuffs the result with dots for alignment.
3492 */
3493 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3494 {
3495 int abblen;
3496 const char *abbrev;
3497 if (len == 40)
3498 return sha1_to_hex(sha1);
3499
3500 abbrev = find_unique_abbrev(sha1, len);
3501 abblen = strlen(abbrev);
3502 if (abblen < 37) {
3503 static char hex[41];
3504 if (len < abblen && abblen <= len + 2)
3505 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3506 else
3507 sprintf(hex, "%s...", abbrev);
3508 return hex;
3509 }
3510 return sha1_to_hex(sha1);
3511 }
3512
3513 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3514 {
3515 int line_termination = opt->line_termination;
3516 int inter_name_termination = line_termination ? '\t' : '\0';
3517 if (opt->output_prefix) {
3518 struct strbuf *msg = NULL;
3519 msg = opt->output_prefix(opt, opt->output_prefix_data);
3520 fprintf(opt->file, "%s", msg->buf);
3521 }
3522
3523 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3524 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3525 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3526 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3527 }
3528 if (p->score) {
3529 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3530 inter_name_termination);
3531 } else {
3532 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3533 }
3534
3535 if (p->status == DIFF_STATUS_COPIED ||
3536 p->status == DIFF_STATUS_RENAMED) {
3537 const char *name_a, *name_b;
3538 name_a = p->one->path;
3539 name_b = p->two->path;
3540 strip_prefix(opt->prefix_length, &name_a, &name_b);
3541 write_name_quoted(name_a, opt->file, inter_name_termination);
3542 write_name_quoted(name_b, opt->file, line_termination);
3543 } else {
3544 const char *name_a, *name_b;
3545 name_a = p->one->mode ? p->one->path : p->two->path;
3546 name_b = NULL;
3547 strip_prefix(opt->prefix_length, &name_a, &name_b);
3548 write_name_quoted(name_a, opt->file, line_termination);
3549 }
3550 }
3551
3552 int diff_unmodified_pair(struct diff_filepair *p)
3553 {
3554 /* This function is written stricter than necessary to support
3555 * the currently implemented transformers, but the idea is to
3556 * let transformers to produce diff_filepairs any way they want,
3557 * and filter and clean them up here before producing the output.
3558 */
3559 struct diff_filespec *one = p->one, *two = p->two;
3560
3561 if (DIFF_PAIR_UNMERGED(p))
3562 return 0; /* unmerged is interesting */
3563
3564 /* deletion, addition, mode or type change
3565 * and rename are all interesting.
3566 */
3567 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3568 DIFF_PAIR_MODE_CHANGED(p) ||
3569 strcmp(one->path, two->path))
3570 return 0;
3571
3572 /* both are valid and point at the same path. that is, we are
3573 * dealing with a change.
3574 */
3575 if (one->sha1_valid && two->sha1_valid &&
3576 !hashcmp(one->sha1, two->sha1) &&
3577 !one->dirty_submodule && !two->dirty_submodule)
3578 return 1; /* no change */
3579 if (!one->sha1_valid && !two->sha1_valid)
3580 return 1; /* both look at the same file on the filesystem. */
3581 return 0;
3582 }
3583
3584 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3585 {
3586 if (diff_unmodified_pair(p))
3587 return;
3588
3589 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3590 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3591 return; /* no tree diffs in patch format */
3592
3593 run_diff(p, o);
3594 }
3595
3596 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3597 struct diffstat_t *diffstat)
3598 {
3599 if (diff_unmodified_pair(p))
3600 return;
3601
3602 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3603 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3604 return; /* no useful stat for tree diffs */
3605
3606 run_diffstat(p, o, diffstat);
3607 }
3608
3609 static void diff_flush_checkdiff(struct diff_filepair *p,
3610 struct diff_options *o)
3611 {
3612 if (diff_unmodified_pair(p))
3613 return;
3614
3615 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3616 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3617 return; /* nothing to check in tree diffs */
3618
3619 run_checkdiff(p, o);
3620 }
3621
3622 int diff_queue_is_empty(void)
3623 {
3624 struct diff_queue_struct *q = &diff_queued_diff;
3625 int i;
3626 for (i = 0; i < q->nr; i++)
3627 if (!diff_unmodified_pair(q->queue[i]))
3628 return 0;
3629 return 1;
3630 }
3631
3632 #if DIFF_DEBUG
3633 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3634 {
3635 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3636 x, one ? one : "",
3637 s->path,
3638 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3639 s->mode,
3640 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3641 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3642 x, one ? one : "",
3643 s->size, s->xfrm_flags);
3644 }
3645
3646 void diff_debug_filepair(const struct diff_filepair *p, int i)
3647 {
3648 diff_debug_filespec(p->one, i, "one");
3649 diff_debug_filespec(p->two, i, "two");
3650 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3651 p->score, p->status ? p->status : '?',
3652 p->one->rename_used, p->broken_pair);
3653 }
3654
3655 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3656 {
3657 int i;
3658 if (msg)
3659 fprintf(stderr, "%s\n", msg);
3660 fprintf(stderr, "q->nr = %d\n", q->nr);
3661 for (i = 0; i < q->nr; i++) {
3662 struct diff_filepair *p = q->queue[i];
3663 diff_debug_filepair(p, i);
3664 }
3665 }
3666 #endif
3667
3668 static void diff_resolve_rename_copy(void)
3669 {
3670 int i;
3671 struct diff_filepair *p;
3672 struct diff_queue_struct *q = &diff_queued_diff;
3673
3674 diff_debug_queue("resolve-rename-copy", q);
3675
3676 for (i = 0; i < q->nr; i++) {
3677 p = q->queue[i];
3678 p->status = 0; /* undecided */
3679 if (DIFF_PAIR_UNMERGED(p))
3680 p->status = DIFF_STATUS_UNMERGED;
3681 else if (!DIFF_FILE_VALID(p->one))
3682 p->status = DIFF_STATUS_ADDED;
3683 else if (!DIFF_FILE_VALID(p->two))
3684 p->status = DIFF_STATUS_DELETED;
3685 else if (DIFF_PAIR_TYPE_CHANGED(p))
3686 p->status = DIFF_STATUS_TYPE_CHANGED;
3687
3688 /* from this point on, we are dealing with a pair
3689 * whose both sides are valid and of the same type, i.e.
3690 * either in-place edit or rename/copy edit.
3691 */
3692 else if (DIFF_PAIR_RENAME(p)) {
3693 /*
3694 * A rename might have re-connected a broken
3695 * pair up, causing the pathnames to be the
3696 * same again. If so, that's not a rename at
3697 * all, just a modification..
3698 *
3699 * Otherwise, see if this source was used for
3700 * multiple renames, in which case we decrement
3701 * the count, and call it a copy.
3702 */
3703 if (!strcmp(p->one->path, p->two->path))
3704 p->status = DIFF_STATUS_MODIFIED;
3705 else if (--p->one->rename_used > 0)
3706 p->status = DIFF_STATUS_COPIED;
3707 else
3708 p->status = DIFF_STATUS_RENAMED;
3709 }
3710 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3711 p->one->mode != p->two->mode ||
3712 p->one->dirty_submodule ||
3713 p->two->dirty_submodule ||
3714 is_null_sha1(p->one->sha1))
3715 p->status = DIFF_STATUS_MODIFIED;
3716 else {
3717 /* This is a "no-change" entry and should not
3718 * happen anymore, but prepare for broken callers.
3719 */
3720 error("feeding unmodified %s to diffcore",
3721 p->one->path);
3722 p->status = DIFF_STATUS_UNKNOWN;
3723 }
3724 }
3725 diff_debug_queue("resolve-rename-copy done", q);
3726 }
3727
3728 static int check_pair_status(struct diff_filepair *p)
3729 {
3730 switch (p->status) {
3731 case DIFF_STATUS_UNKNOWN:
3732 return 0;
3733 case 0:
3734 die("internal error in diff-resolve-rename-copy");
3735 default:
3736 return 1;
3737 }
3738 }
3739
3740 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3741 {
3742 int fmt = opt->output_format;
3743
3744 if (fmt & DIFF_FORMAT_CHECKDIFF)
3745 diff_flush_checkdiff(p, opt);
3746 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3747 diff_flush_raw(p, opt);
3748 else if (fmt & DIFF_FORMAT_NAME) {
3749 const char *name_a, *name_b;
3750 name_a = p->two->path;
3751 name_b = NULL;
3752 strip_prefix(opt->prefix_length, &name_a, &name_b);
3753 write_name_quoted(name_a, opt->file, opt->line_termination);
3754 }
3755 }
3756
3757 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3758 {
3759 if (fs->mode)
3760 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3761 else
3762 fprintf(file, " %s ", newdelete);
3763 write_name_quoted(fs->path, file, '\n');
3764 }
3765
3766
3767 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3768 const char *line_prefix)
3769 {
3770 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3771 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3772 p->two->mode, show_name ? ' ' : '\n');
3773 if (show_name) {
3774 write_name_quoted(p->two->path, file, '\n');
3775 }
3776 }
3777 }
3778
3779 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3780 const char *line_prefix)
3781 {
3782 char *names = pprint_rename(p->one->path, p->two->path);
3783
3784 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3785 free(names);
3786 show_mode_change(file, p, 0, line_prefix);
3787 }
3788
3789 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3790 {
3791 FILE *file = opt->file;
3792 char *line_prefix = "";
3793
3794 if (opt->output_prefix) {
3795 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3796 line_prefix = buf->buf;
3797 }
3798
3799 switch(p->status) {
3800 case DIFF_STATUS_DELETED:
3801 fputs(line_prefix, file);
3802 show_file_mode_name(file, "delete", p->one);
3803 break;
3804 case DIFF_STATUS_ADDED:
3805 fputs(line_prefix, file);
3806 show_file_mode_name(file, "create", p->two);
3807 break;
3808 case DIFF_STATUS_COPIED:
3809 fputs(line_prefix, file);
3810 show_rename_copy(file, "copy", p, line_prefix);
3811 break;
3812 case DIFF_STATUS_RENAMED:
3813 fputs(line_prefix, file);
3814 show_rename_copy(file, "rename", p, line_prefix);
3815 break;
3816 default:
3817 if (p->score) {
3818 fprintf(file, "%s rewrite ", line_prefix);
3819 write_name_quoted(p->two->path, file, ' ');
3820 fprintf(file, "(%d%%)\n", similarity_index(p));
3821 }
3822 show_mode_change(file, p, !p->score, line_prefix);
3823 break;
3824 }
3825 }
3826
3827 struct patch_id_t {
3828 git_SHA_CTX *ctx;
3829 int patchlen;
3830 };
3831
3832 static int remove_space(char *line, int len)
3833 {
3834 int i;
3835 char *dst = line;
3836 unsigned char c;
3837
3838 for (i = 0; i < len; i++)
3839 if (!isspace((c = line[i])))
3840 *dst++ = c;
3841
3842 return dst - line;
3843 }
3844
3845 static void patch_id_consume(void *priv, char *line, unsigned long len)
3846 {
3847 struct patch_id_t *data = priv;
3848 int new_len;
3849
3850 /* Ignore line numbers when computing the SHA1 of the patch */
3851 if (!prefixcmp(line, "@@ -"))
3852 return;
3853
3854 new_len = remove_space(line, len);
3855
3856 git_SHA1_Update(data->ctx, line, new_len);
3857 data->patchlen += new_len;
3858 }
3859
3860 /* returns 0 upon success, and writes result into sha1 */
3861 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3862 {
3863 struct diff_queue_struct *q = &diff_queued_diff;
3864 int i;
3865 git_SHA_CTX ctx;
3866 struct patch_id_t data;
3867 char buffer[PATH_MAX * 4 + 20];
3868
3869 git_SHA1_Init(&ctx);
3870 memset(&data, 0, sizeof(struct patch_id_t));
3871 data.ctx = &ctx;
3872
3873 for (i = 0; i < q->nr; i++) {
3874 xpparam_t xpp;
3875 xdemitconf_t xecfg;
3876 mmfile_t mf1, mf2;
3877 struct diff_filepair *p = q->queue[i];
3878 int len1, len2;
3879
3880 memset(&xpp, 0, sizeof(xpp));
3881 memset(&xecfg, 0, sizeof(xecfg));
3882 if (p->status == 0)
3883 return error("internal diff status error");
3884 if (p->status == DIFF_STATUS_UNKNOWN)
3885 continue;
3886 if (diff_unmodified_pair(p))
3887 continue;
3888 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3889 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3890 continue;
3891 if (DIFF_PAIR_UNMERGED(p))
3892 continue;
3893
3894 diff_fill_sha1_info(p->one);
3895 diff_fill_sha1_info(p->two);
3896 if (fill_mmfile(&mf1, p->one) < 0 ||
3897 fill_mmfile(&mf2, p->two) < 0)
3898 return error("unable to read files to diff");
3899
3900 len1 = remove_space(p->one->path, strlen(p->one->path));
3901 len2 = remove_space(p->two->path, strlen(p->two->path));
3902 if (p->one->mode == 0)
3903 len1 = snprintf(buffer, sizeof(buffer),
3904 "diff--gita/%.*sb/%.*s"
3905 "newfilemode%06o"
3906 "---/dev/null"
3907 "+++b/%.*s",
3908 len1, p->one->path,
3909 len2, p->two->path,
3910 p->two->mode,
3911 len2, p->two->path);
3912 else if (p->two->mode == 0)
3913 len1 = snprintf(buffer, sizeof(buffer),
3914 "diff--gita/%.*sb/%.*s"
3915 "deletedfilemode%06o"
3916 "---a/%.*s"
3917 "+++/dev/null",
3918 len1, p->one->path,
3919 len2, p->two->path,
3920 p->one->mode,
3921 len1, p->one->path);
3922 else
3923 len1 = snprintf(buffer, sizeof(buffer),
3924 "diff--gita/%.*sb/%.*s"
3925 "---a/%.*s"
3926 "+++b/%.*s",
3927 len1, p->one->path,
3928 len2, p->two->path,
3929 len1, p->one->path,
3930 len2, p->two->path);
3931 git_SHA1_Update(&ctx, buffer, len1);
3932
3933 if (diff_filespec_is_binary(p->one) ||
3934 diff_filespec_is_binary(p->two)) {
3935 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3936 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3937 continue;
3938 }
3939
3940 xpp.flags = 0;
3941 xecfg.ctxlen = 3;
3942 xecfg.flags = 0;
3943 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3944 &xpp, &xecfg);
3945 }
3946
3947 git_SHA1_Final(sha1, &ctx);
3948 return 0;
3949 }
3950
3951 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3952 {
3953 struct diff_queue_struct *q = &diff_queued_diff;
3954 int i;
3955 int result = diff_get_patch_id(options, sha1);
3956
3957 for (i = 0; i < q->nr; i++)
3958 diff_free_filepair(q->queue[i]);
3959
3960 free(q->queue);
3961 DIFF_QUEUE_CLEAR(q);
3962
3963 return result;
3964 }
3965
3966 static int is_summary_empty(const struct diff_queue_struct *q)
3967 {
3968 int i;
3969
3970 for (i = 0; i < q->nr; i++) {
3971 const struct diff_filepair *p = q->queue[i];
3972
3973 switch (p->status) {
3974 case DIFF_STATUS_DELETED:
3975 case DIFF_STATUS_ADDED:
3976 case DIFF_STATUS_COPIED:
3977 case DIFF_STATUS_RENAMED:
3978 return 0;
3979 default:
3980 if (p->score)
3981 return 0;
3982 if (p->one->mode && p->two->mode &&
3983 p->one->mode != p->two->mode)
3984 return 0;
3985 break;
3986 }
3987 }
3988 return 1;
3989 }
3990
3991 void diff_flush(struct diff_options *options)
3992 {
3993 struct diff_queue_struct *q = &diff_queued_diff;
3994 int i, output_format = options->output_format;
3995 int separator = 0;
3996
3997 /*
3998 * Order: raw, stat, summary, patch
3999 * or: name/name-status/checkdiff (other bits clear)
4000 */
4001 if (!q->nr)
4002 goto free_queue;
4003
4004 if (output_format & (DIFF_FORMAT_RAW |
4005 DIFF_FORMAT_NAME |
4006 DIFF_FORMAT_NAME_STATUS |
4007 DIFF_FORMAT_CHECKDIFF)) {
4008 for (i = 0; i < q->nr; i++) {
4009 struct diff_filepair *p = q->queue[i];
4010 if (check_pair_status(p))
4011 flush_one_pair(p, options);
4012 }
4013 separator++;
4014 }
4015
4016 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4017 struct diffstat_t diffstat;
4018
4019 memset(&diffstat, 0, sizeof(struct diffstat_t));
4020 for (i = 0; i < q->nr; i++) {
4021 struct diff_filepair *p = q->queue[i];
4022 if (check_pair_status(p))
4023 diff_flush_stat(p, options, &diffstat);
4024 }
4025 if (output_format & DIFF_FORMAT_NUMSTAT)
4026 show_numstat(&diffstat, options);
4027 if (output_format & DIFF_FORMAT_DIFFSTAT)
4028 show_stats(&diffstat, options);
4029 if (output_format & DIFF_FORMAT_SHORTSTAT)
4030 show_shortstats(&diffstat, options);
4031 free_diffstat_info(&diffstat);
4032 separator++;
4033 }
4034 if (output_format & DIFF_FORMAT_DIRSTAT)
4035 show_dirstat(options);
4036
4037 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4038 for (i = 0; i < q->nr; i++) {
4039 diff_summary(options, q->queue[i]);
4040 }
4041 separator++;
4042 }
4043
4044 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4045 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4046 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4047 /*
4048 * run diff_flush_patch for the exit status. setting
4049 * options->file to /dev/null should be safe, becaue we
4050 * aren't supposed to produce any output anyway.
4051 */
4052 if (options->close_file)
4053 fclose(options->file);
4054 options->file = fopen("/dev/null", "w");
4055 if (!options->file)
4056 die_errno("Could not open /dev/null");
4057 options->close_file = 1;
4058 for (i = 0; i < q->nr; i++) {
4059 struct diff_filepair *p = q->queue[i];
4060 if (check_pair_status(p))
4061 diff_flush_patch(p, options);
4062 if (options->found_changes)
4063 break;
4064 }
4065 }
4066
4067 if (output_format & DIFF_FORMAT_PATCH) {
4068 if (separator) {
4069 putc(options->line_termination, options->file);
4070 if (options->stat_sep) {
4071 /* attach patch instead of inline */
4072 fputs(options->stat_sep, options->file);
4073 }
4074 }
4075
4076 for (i = 0; i < q->nr; i++) {
4077 struct diff_filepair *p = q->queue[i];
4078 if (check_pair_status(p))
4079 diff_flush_patch(p, options);
4080 }
4081 }
4082
4083 if (output_format & DIFF_FORMAT_CALLBACK)
4084 options->format_callback(q, options, options->format_callback_data);
4085
4086 for (i = 0; i < q->nr; i++)
4087 diff_free_filepair(q->queue[i]);
4088 free_queue:
4089 free(q->queue);
4090 DIFF_QUEUE_CLEAR(q);
4091 if (options->close_file)
4092 fclose(options->file);
4093
4094 /*
4095 * Report the content-level differences with HAS_CHANGES;
4096 * diff_addremove/diff_change does not set the bit when
4097 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4098 */
4099 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4100 if (options->found_changes)
4101 DIFF_OPT_SET(options, HAS_CHANGES);
4102 else
4103 DIFF_OPT_CLR(options, HAS_CHANGES);
4104 }
4105 }
4106
4107 static void diffcore_apply_filter(const char *filter)
4108 {
4109 int i;
4110 struct diff_queue_struct *q = &diff_queued_diff;
4111 struct diff_queue_struct outq;
4112 DIFF_QUEUE_CLEAR(&outq);
4113
4114 if (!filter)
4115 return;
4116
4117 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4118 int found;
4119 for (i = found = 0; !found && i < q->nr; i++) {
4120 struct diff_filepair *p = q->queue[i];
4121 if (((p->status == DIFF_STATUS_MODIFIED) &&
4122 ((p->score &&
4123 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4124 (!p->score &&
4125 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4126 ((p->status != DIFF_STATUS_MODIFIED) &&
4127 strchr(filter, p->status)))
4128 found++;
4129 }
4130 if (found)
4131 return;
4132
4133 /* otherwise we will clear the whole queue
4134 * by copying the empty outq at the end of this
4135 * function, but first clear the current entries
4136 * in the queue.
4137 */
4138 for (i = 0; i < q->nr; i++)
4139 diff_free_filepair(q->queue[i]);
4140 }
4141 else {
4142 /* Only the matching ones */
4143 for (i = 0; i < q->nr; i++) {
4144 struct diff_filepair *p = q->queue[i];
4145
4146 if (((p->status == DIFF_STATUS_MODIFIED) &&
4147 ((p->score &&
4148 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4149 (!p->score &&
4150 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4151 ((p->status != DIFF_STATUS_MODIFIED) &&
4152 strchr(filter, p->status)))
4153 diff_q(&outq, p);
4154 else
4155 diff_free_filepair(p);
4156 }
4157 }
4158 free(q->queue);
4159 *q = outq;
4160 }
4161
4162 /* Check whether two filespecs with the same mode and size are identical */
4163 static int diff_filespec_is_identical(struct diff_filespec *one,
4164 struct diff_filespec *two)
4165 {
4166 if (S_ISGITLINK(one->mode))
4167 return 0;
4168 if (diff_populate_filespec(one, 0))
4169 return 0;
4170 if (diff_populate_filespec(two, 0))
4171 return 0;
4172 return !memcmp(one->data, two->data, one->size);
4173 }
4174
4175 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4176 {
4177 int i;
4178 struct diff_queue_struct *q = &diff_queued_diff;
4179 struct diff_queue_struct outq;
4180 DIFF_QUEUE_CLEAR(&outq);
4181
4182 for (i = 0; i < q->nr; i++) {
4183 struct diff_filepair *p = q->queue[i];
4184
4185 /*
4186 * 1. Entries that come from stat info dirtiness
4187 * always have both sides (iow, not create/delete),
4188 * one side of the object name is unknown, with
4189 * the same mode and size. Keep the ones that
4190 * do not match these criteria. They have real
4191 * differences.
4192 *
4193 * 2. At this point, the file is known to be modified,
4194 * with the same mode and size, and the object
4195 * name of one side is unknown. Need to inspect
4196 * the identical contents.
4197 */
4198 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4199 !DIFF_FILE_VALID(p->two) ||
4200 (p->one->sha1_valid && p->two->sha1_valid) ||
4201 (p->one->mode != p->two->mode) ||
4202 diff_populate_filespec(p->one, 1) ||
4203 diff_populate_filespec(p->two, 1) ||
4204 (p->one->size != p->two->size) ||
4205 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4206 diff_q(&outq, p);
4207 else {
4208 /*
4209 * The caller can subtract 1 from skip_stat_unmatch
4210 * to determine how many paths were dirty only
4211 * due to stat info mismatch.
4212 */
4213 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4214 diffopt->skip_stat_unmatch++;
4215 diff_free_filepair(p);
4216 }
4217 }
4218 free(q->queue);
4219 *q = outq;
4220 }
4221
4222 static int diffnamecmp(const void *a_, const void *b_)
4223 {
4224 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4225 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4226 const char *name_a, *name_b;
4227
4228 name_a = a->one ? a->one->path : a->two->path;
4229 name_b = b->one ? b->one->path : b->two->path;
4230 return strcmp(name_a, name_b);
4231 }
4232
4233 void diffcore_fix_diff_index(struct diff_options *options)
4234 {
4235 struct diff_queue_struct *q = &diff_queued_diff;
4236 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4237 }
4238
4239 void diffcore_std(struct diff_options *options)
4240 {
4241 if (options->skip_stat_unmatch)
4242 diffcore_skip_stat_unmatch(options);
4243 if (!options->found_follow) {
4244 /* See try_to_follow_renames() in tree-diff.c */
4245 if (options->break_opt != -1)
4246 diffcore_break(options->break_opt);
4247 if (options->detect_rename)
4248 diffcore_rename(options);
4249 if (options->break_opt != -1)
4250 diffcore_merge_broken();
4251 }
4252 if (options->pickaxe)
4253 diffcore_pickaxe(options);
4254 if (options->orderfile)
4255 diffcore_order(options->orderfile);
4256 if (!options->found_follow)
4257 /* See try_to_follow_renames() in tree-diff.c */
4258 diff_resolve_rename_copy();
4259 diffcore_apply_filter(options->filter);
4260
4261 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4262 DIFF_OPT_SET(options, HAS_CHANGES);
4263 else
4264 DIFF_OPT_CLR(options, HAS_CHANGES);
4265
4266 options->found_follow = 0;
4267 }
4268
4269 int diff_result_code(struct diff_options *opt, int status)
4270 {
4271 int result = 0;
4272 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4273 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4274 return status;
4275 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4276 DIFF_OPT_TST(opt, HAS_CHANGES))
4277 result |= 01;
4278 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4279 DIFF_OPT_TST(opt, CHECK_FAILED))
4280 result |= 02;
4281 return result;
4282 }
4283
4284 /*
4285 * Shall changes to this submodule be ignored?
4286 *
4287 * Submodule changes can be configured to be ignored separately for each path,
4288 * but that configuration can be overridden from the command line.
4289 */
4290 static int is_submodule_ignored(const char *path, struct diff_options *options)
4291 {
4292 int ignored = 0;
4293 unsigned orig_flags = options->flags;
4294 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4295 set_diffopt_flags_from_submodule_config(options, path);
4296 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4297 ignored = 1;
4298 options->flags = orig_flags;
4299 return ignored;
4300 }
4301
4302 void diff_addremove(struct diff_options *options,
4303 int addremove, unsigned mode,
4304 const unsigned char *sha1,
4305 const char *concatpath, unsigned dirty_submodule)
4306 {
4307 struct diff_filespec *one, *two;
4308
4309 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4310 return;
4311
4312 /* This may look odd, but it is a preparation for
4313 * feeding "there are unchanged files which should
4314 * not produce diffs, but when you are doing copy
4315 * detection you would need them, so here they are"
4316 * entries to the diff-core. They will be prefixed
4317 * with something like '=' or '*' (I haven't decided
4318 * which but should not make any difference).
4319 * Feeding the same new and old to diff_change()
4320 * also has the same effect.
4321 * Before the final output happens, they are pruned after
4322 * merged into rename/copy pairs as appropriate.
4323 */
4324 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4325 addremove = (addremove == '+' ? '-' :
4326 addremove == '-' ? '+' : addremove);
4327
4328 if (options->prefix &&
4329 strncmp(concatpath, options->prefix, options->prefix_length))
4330 return;
4331
4332 one = alloc_filespec(concatpath);
4333 two = alloc_filespec(concatpath);
4334
4335 if (addremove != '+')
4336 fill_filespec(one, sha1, mode);
4337 if (addremove != '-') {
4338 fill_filespec(two, sha1, mode);
4339 two->dirty_submodule = dirty_submodule;
4340 }
4341
4342 diff_queue(&diff_queued_diff, one, two);
4343 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4344 DIFF_OPT_SET(options, HAS_CHANGES);
4345 }
4346
4347 void diff_change(struct diff_options *options,
4348 unsigned old_mode, unsigned new_mode,
4349 const unsigned char *old_sha1,
4350 const unsigned char *new_sha1,
4351 const char *concatpath,
4352 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4353 {
4354 struct diff_filespec *one, *two;
4355
4356 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4357 is_submodule_ignored(concatpath, options))
4358 return;
4359
4360 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4361 unsigned tmp;
4362 const unsigned char *tmp_c;
4363 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4364 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4365 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4366 new_dirty_submodule = tmp;
4367 }
4368
4369 if (options->prefix &&
4370 strncmp(concatpath, options->prefix, options->prefix_length))
4371 return;
4372
4373 one = alloc_filespec(concatpath);
4374 two = alloc_filespec(concatpath);
4375 fill_filespec(one, old_sha1, old_mode);
4376 fill_filespec(two, new_sha1, new_mode);
4377 one->dirty_submodule = old_dirty_submodule;
4378 two->dirty_submodule = new_dirty_submodule;
4379
4380 diff_queue(&diff_queued_diff, one, two);
4381 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4382 DIFF_OPT_SET(options, HAS_CHANGES);
4383 }
4384
4385 void diff_unmerge(struct diff_options *options,
4386 const char *path,
4387 unsigned mode, const unsigned char *sha1)
4388 {
4389 struct diff_filespec *one, *two;
4390
4391 if (options->prefix &&
4392 strncmp(path, options->prefix, options->prefix_length))
4393 return;
4394
4395 one = alloc_filespec(path);
4396 two = alloc_filespec(path);
4397 fill_filespec(one, sha1, mode);
4398 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4399 }
4400
4401 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4402 size_t *outsize)
4403 {
4404 struct diff_tempfile *temp;
4405 const char *argv[3];
4406 const char **arg = argv;
4407 struct child_process child;
4408 struct strbuf buf = STRBUF_INIT;
4409 int err = 0;
4410
4411 temp = prepare_temp_file(spec->path, spec);
4412 *arg++ = pgm;
4413 *arg++ = temp->name;
4414 *arg = NULL;
4415
4416 memset(&child, 0, sizeof(child));
4417 child.use_shell = 1;
4418 child.argv = argv;
4419 child.out = -1;
4420 if (start_command(&child)) {
4421 remove_tempfile();
4422 return NULL;
4423 }
4424
4425 if (strbuf_read(&buf, child.out, 0) < 0)
4426 err = error("error reading from textconv command '%s'", pgm);
4427 close(child.out);
4428
4429 if (finish_command(&child) || err) {
4430 strbuf_release(&buf);
4431 remove_tempfile();
4432 return NULL;
4433 }
4434 remove_tempfile();
4435
4436 return strbuf_detach(&buf, outsize);
4437 }
4438
4439 size_t fill_textconv(struct userdiff_driver *driver,
4440 struct diff_filespec *df,
4441 char **outbuf)
4442 {
4443 size_t size;
4444
4445 if (!driver || !driver->textconv) {
4446 if (!DIFF_FILE_VALID(df)) {
4447 *outbuf = "";
4448 return 0;
4449 }
4450 if (diff_populate_filespec(df, 0))
4451 die("unable to read files to diff");
4452 *outbuf = df->data;
4453 return df->size;
4454 }
4455
4456 if (driver->textconv_cache && df->sha1_valid) {
4457 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4458 &size);
4459 if (*outbuf)
4460 return size;
4461 }
4462
4463 *outbuf = run_textconv(driver->textconv, df, &size);
4464 if (!*outbuf)
4465 die("unable to read files to diff");
4466
4467 if (driver->textconv_cache && df->sha1_valid) {
4468 /* ignore errors, as we might be in a readonly repository */
4469 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4470 size);
4471 /*
4472 * we could save up changes and flush them all at the end,
4473 * but we would need an extra call after all diffing is done.
4474 * Since generating a cache entry is the slow path anyway,
4475 * this extra overhead probably isn't a big deal.
4476 */
4477 notes_cache_write(driver->textconv_cache);
4478 }
4479
4480 return size;
4481 }
Jonathan Nieder’s miscellaneous git branches