?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff --stat-count: finishing touches
[git/jrn.git] / diff.c
blob542ff42e5aa28b2a235466e07795bd2e6214afd7
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 400;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static struct diff_options default_diff_options;
35
36 static char diff_colors[][COLOR_MAXLEN] = {
37 GIT_COLOR_RESET,
38 GIT_COLOR_NORMAL, /* PLAIN */
39 GIT_COLOR_BOLD, /* METAINFO */
40 GIT_COLOR_CYAN, /* FRAGINFO */
41 GIT_COLOR_RED, /* OLD */
42 GIT_COLOR_GREEN, /* NEW */
43 GIT_COLOR_YELLOW, /* COMMIT */
44 GIT_COLOR_BG_RED, /* WHITESPACE */
45 GIT_COLOR_NORMAL, /* FUNCINFO */
46 };
47
48 static int parse_diff_color_slot(const char *var, int ofs)
49 {
50 if (!strcasecmp(var+ofs, "plain"))
51 return DIFF_PLAIN;
52 if (!strcasecmp(var+ofs, "meta"))
53 return DIFF_METAINFO;
54 if (!strcasecmp(var+ofs, "frag"))
55 return DIFF_FRAGINFO;
56 if (!strcasecmp(var+ofs, "old"))
57 return DIFF_FILE_OLD;
58 if (!strcasecmp(var+ofs, "new"))
59 return DIFF_FILE_NEW;
60 if (!strcasecmp(var+ofs, "commit"))
61 return DIFF_COMMIT;
62 if (!strcasecmp(var+ofs, "whitespace"))
63 return DIFF_WHITESPACE;
64 if (!strcasecmp(var+ofs, "func"))
65 return DIFF_FUNCINFO;
66 return -1;
67 }
68
69 static int git_config_rename(const char *var, const char *value)
70 {
71 if (!value)
72 return DIFF_DETECT_RENAME;
73 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
74 return DIFF_DETECT_COPY;
75 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
76 }
77
78 /*
79 * These are to give UI layer defaults.
80 * The core-level commands such as git-diff-files should
81 * never be affected by the setting of diff.renames
82 * the user happens to have in the configuration file.
83 */
84 int git_diff_ui_config(const char *var, const char *value, void *cb)
85 {
86 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
87 diff_use_color_default = git_config_colorbool(var, value, -1);
88 return 0;
89 }
90 if (!strcmp(var, "diff.renames")) {
91 diff_detect_rename_default = git_config_rename(var, value);
92 return 0;
93 }
94 if (!strcmp(var, "diff.autorefreshindex")) {
95 diff_auto_refresh_index = git_config_bool(var, value);
96 return 0;
97 }
98 if (!strcmp(var, "diff.mnemonicprefix")) {
99 diff_mnemonic_prefix = git_config_bool(var, value);
100 return 0;
101 }
102 if (!strcmp(var, "diff.noprefix")) {
103 diff_no_prefix = git_config_bool(var, value);
104 return 0;
105 }
106 if (!strcmp(var, "diff.external"))
107 return git_config_string(&external_diff_cmd_cfg, var, value);
108 if (!strcmp(var, "diff.wordregex"))
109 return git_config_string(&diff_word_regex_cfg, var, value);
110
111 if (!strcmp(var, "diff.ignoresubmodules"))
112 handle_ignore_submodules_arg(&default_diff_options, value);
113
114 return git_diff_basic_config(var, value, cb);
115 }
116
117 int git_diff_basic_config(const char *var, const char *value, void *cb)
118 {
119 if (!strcmp(var, "diff.renamelimit")) {
120 diff_rename_limit_default = git_config_int(var, value);
121 return 0;
122 }
123
124 switch (userdiff_config(var, value)) {
125 case 0: break;
126 case -1: return -1;
127 default: return 0;
128 }
129
130 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
131 int slot = parse_diff_color_slot(var, 11);
132 if (slot < 0)
133 return 0;
134 if (!value)
135 return config_error_nonbool(var);
136 color_parse(value, var, diff_colors[slot]);
137 return 0;
138 }
139
140 /* like GNU diff's --suppress-blank-empty option */
141 if (!strcmp(var, "diff.suppressblankempty") ||
142 /* for backwards compatibility */
143 !strcmp(var, "diff.suppress-blank-empty")) {
144 diff_suppress_blank_empty = git_config_bool(var, value);
145 return 0;
146 }
147
148 if (!prefixcmp(var, "submodule."))
149 return parse_submodule_config_option(var, value);
150
151 return git_color_default_config(var, value, cb);
152 }
153
154 static char *quote_two(const char *one, const char *two)
155 {
156 int need_one = quote_c_style(one, NULL, NULL, 1);
157 int need_two = quote_c_style(two, NULL, NULL, 1);
158 struct strbuf res = STRBUF_INIT;
159
160 if (need_one + need_two) {
161 strbuf_addch(&res, '"');
162 quote_c_style(one, &res, NULL, 1);
163 quote_c_style(two, &res, NULL, 1);
164 strbuf_addch(&res, '"');
165 } else {
166 strbuf_addstr(&res, one);
167 strbuf_addstr(&res, two);
168 }
169 return strbuf_detach(&res, NULL);
170 }
171
172 static const char *external_diff(void)
173 {
174 static const char *external_diff_cmd = NULL;
175 static int done_preparing = 0;
176
177 if (done_preparing)
178 return external_diff_cmd;
179 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
180 if (!external_diff_cmd)
181 external_diff_cmd = external_diff_cmd_cfg;
182 done_preparing = 1;
183 return external_diff_cmd;
184 }
185
186 static struct diff_tempfile {
187 const char *name; /* filename external diff should read from */
188 char hex[41];
189 char mode[10];
190 char tmp_path[PATH_MAX];
191 } diff_temp[2];
192
193 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
194
195 struct emit_callback {
196 int color_diff;
197 unsigned ws_rule;
198 int blank_at_eof_in_preimage;
199 int blank_at_eof_in_postimage;
200 int lno_in_preimage;
201 int lno_in_postimage;
202 sane_truncate_fn truncate;
203 const char **label_path;
204 struct diff_words_data *diff_words;
205 struct diff_options *opt;
206 int *found_changesp;
207 struct strbuf *header;
208 };
209
210 static int count_lines(const char *data, int size)
211 {
212 int count, ch, completely_empty = 1, nl_just_seen = 0;
213 count = 0;
214 while (0 < size--) {
215 ch = *data++;
216 if (ch == '\n') {
217 count++;
218 nl_just_seen = 1;
219 completely_empty = 0;
220 }
221 else {
222 nl_just_seen = 0;
223 completely_empty = 0;
224 }
225 }
226 if (completely_empty)
227 return 0;
228 if (!nl_just_seen)
229 count++; /* no trailing newline */
230 return count;
231 }
232
233 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
234 {
235 if (!DIFF_FILE_VALID(one)) {
236 mf->ptr = (char *)""; /* does not matter */
237 mf->size = 0;
238 return 0;
239 }
240 else if (diff_populate_filespec(one, 0))
241 return -1;
242
243 mf->ptr = one->data;
244 mf->size = one->size;
245 return 0;
246 }
247
248 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
249 static unsigned long diff_filespec_size(struct diff_filespec *one)
250 {
251 if (!DIFF_FILE_VALID(one))
252 return 0;
253 diff_populate_filespec(one, 1);
254 return one->size;
255 }
256
257 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
258 {
259 char *ptr = mf->ptr;
260 long size = mf->size;
261 int cnt = 0;
262
263 if (!size)
264 return cnt;
265 ptr += size - 1; /* pointing at the very end */
266 if (*ptr != '\n')
267 ; /* incomplete line */
268 else
269 ptr--; /* skip the last LF */
270 while (mf->ptr < ptr) {
271 char *prev_eol;
272 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
273 if (*prev_eol == '\n')
274 break;
275 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
276 break;
277 cnt++;
278 ptr = prev_eol - 1;
279 }
280 return cnt;
281 }
282
283 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
284 struct emit_callback *ecbdata)
285 {
286 int l1, l2, at;
287 unsigned ws_rule = ecbdata->ws_rule;
288 l1 = count_trailing_blank(mf1, ws_rule);
289 l2 = count_trailing_blank(mf2, ws_rule);
290 if (l2 <= l1) {
291 ecbdata->blank_at_eof_in_preimage = 0;
292 ecbdata->blank_at_eof_in_postimage = 0;
293 return;
294 }
295 at = count_lines(mf1->ptr, mf1->size);
296 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
297
298 at = count_lines(mf2->ptr, mf2->size);
299 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
300 }
301
302 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
303 int first, const char *line, int len)
304 {
305 int has_trailing_newline, has_trailing_carriage_return;
306 int nofirst;
307 FILE *file = o->file;
308
309 if (o->output_prefix) {
310 struct strbuf *msg = NULL;
311 msg = o->output_prefix(o, o->output_prefix_data);
312 assert(msg);
313 fwrite(msg->buf, msg->len, 1, file);
314 }
315
316 if (len == 0) {
317 has_trailing_newline = (first == '\n');
318 has_trailing_carriage_return = (!has_trailing_newline &&
319 (first == '\r'));
320 nofirst = has_trailing_newline || has_trailing_carriage_return;
321 } else {
322 has_trailing_newline = (len > 0 && line[len-1] == '\n');
323 if (has_trailing_newline)
324 len--;
325 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
326 if (has_trailing_carriage_return)
327 len--;
328 nofirst = 0;
329 }
330
331 if (len || !nofirst) {
332 fputs(set, file);
333 if (!nofirst)
334 fputc(first, file);
335 fwrite(line, len, 1, file);
336 fputs(reset, file);
337 }
338 if (has_trailing_carriage_return)
339 fputc('\r', file);
340 if (has_trailing_newline)
341 fputc('\n', file);
342 }
343
344 static void emit_line(struct diff_options *o, const char *set, const char *reset,
345 const char *line, int len)
346 {
347 emit_line_0(o, set, reset, line[0], line+1, len-1);
348 }
349
350 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
351 {
352 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
353 ecbdata->blank_at_eof_in_preimage &&
354 ecbdata->blank_at_eof_in_postimage &&
355 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
356 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
357 return 0;
358 return ws_blank_line(line, len, ecbdata->ws_rule);
359 }
360
361 static void emit_add_line(const char *reset,
362 struct emit_callback *ecbdata,
363 const char *line, int len)
364 {
365 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
366 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
367
368 if (!*ws)
369 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
370 else if (new_blank_line_at_eof(ecbdata, line, len))
371 /* Blank line at EOF - paint '+' as well */
372 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
373 else {
374 /* Emit just the prefix, then the rest. */
375 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
376 ws_check_emit(line, len, ecbdata->ws_rule,
377 ecbdata->opt->file, set, reset, ws);
378 }
379 }
380
381 static void emit_hunk_header(struct emit_callback *ecbdata,
382 const char *line, int len)
383 {
384 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
385 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
386 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
387 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
388 static const char atat[2] = { '@', '@' };
389 const char *cp, *ep;
390 struct strbuf msgbuf = STRBUF_INIT;
391 int org_len = len;
392 int i = 1;
393
394 /*
395 * As a hunk header must begin with "@@ -<old>, +<new> @@",
396 * it always is at least 10 bytes long.
397 */
398 if (len < 10 ||
399 memcmp(line, atat, 2) ||
400 !(ep = memmem(line + 2, len - 2, atat, 2))) {
401 emit_line(ecbdata->opt, plain, reset, line, len);
402 return;
403 }
404 ep += 2; /* skip over @@ */
405
406 /* The hunk header in fraginfo color */
407 strbuf_add(&msgbuf, frag, strlen(frag));
408 strbuf_add(&msgbuf, line, ep - line);
409 strbuf_add(&msgbuf, reset, strlen(reset));
410
411 /*
412 * trailing "\r\n"
413 */
414 for ( ; i < 3; i++)
415 if (line[len - i] == '\r' || line[len - i] == '\n')
416 len--;
417
418 /* blank before the func header */
419 for (cp = ep; ep - line < len; ep++)
420 if (*ep != ' ' && *ep != '\t')
421 break;
422 if (ep != cp) {
423 strbuf_add(&msgbuf, plain, strlen(plain));
424 strbuf_add(&msgbuf, cp, ep - cp);
425 strbuf_add(&msgbuf, reset, strlen(reset));
426 }
427
428 if (ep < line + len) {
429 strbuf_add(&msgbuf, func, strlen(func));
430 strbuf_add(&msgbuf, ep, line + len - ep);
431 strbuf_add(&msgbuf, reset, strlen(reset));
432 }
433
434 strbuf_add(&msgbuf, line + len, org_len - len);
435 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
436 strbuf_release(&msgbuf);
437 }
438
439 static struct diff_tempfile *claim_diff_tempfile(void) {
440 int i;
441 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
442 if (!diff_temp[i].name)
443 return diff_temp + i;
444 die("BUG: diff is failing to clean up its tempfiles");
445 }
446
447 static int remove_tempfile_installed;
448
449 static void remove_tempfile(void)
450 {
451 int i;
452 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
453 if (diff_temp[i].name == diff_temp[i].tmp_path)
454 unlink_or_warn(diff_temp[i].name);
455 diff_temp[i].name = NULL;
456 }
457 }
458
459 static void remove_tempfile_on_signal(int signo)
460 {
461 remove_tempfile();
462 sigchain_pop(signo);
463 raise(signo);
464 }
465
466 static void print_line_count(FILE *file, int count)
467 {
468 switch (count) {
469 case 0:
470 fprintf(file, "0,0");
471 break;
472 case 1:
473 fprintf(file, "1");
474 break;
475 default:
476 fprintf(file, "1,%d", count);
477 break;
478 }
479 }
480
481 static void emit_rewrite_lines(struct emit_callback *ecb,
482 int prefix, const char *data, int size)
483 {
484 const char *endp = NULL;
485 static const char *nneof = " No newline at end of file\n";
486 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
487 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
488
489 while (0 < size) {
490 int len;
491
492 endp = memchr(data, '\n', size);
493 len = endp ? (endp - data + 1) : size;
494 if (prefix != '+') {
495 ecb->lno_in_preimage++;
496 emit_line_0(ecb->opt, old, reset, '-',
497 data, len);
498 } else {
499 ecb->lno_in_postimage++;
500 emit_add_line(reset, ecb, data, len);
501 }
502 size -= len;
503 data += len;
504 }
505 if (!endp) {
506 const char *plain = diff_get_color(ecb->color_diff,
507 DIFF_PLAIN);
508 emit_line_0(ecb->opt, plain, reset, '\\',
509 nneof, strlen(nneof));
510 }
511 }
512
513 static void emit_rewrite_diff(const char *name_a,
514 const char *name_b,
515 struct diff_filespec *one,
516 struct diff_filespec *two,
517 struct userdiff_driver *textconv_one,
518 struct userdiff_driver *textconv_two,
519 struct diff_options *o)
520 {
521 int lc_a, lc_b;
522 int color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
523 const char *name_a_tab, *name_b_tab;
524 const char *metainfo = diff_get_color(color_diff, DIFF_METAINFO);
525 const char *fraginfo = diff_get_color(color_diff, DIFF_FRAGINFO);
526 const char *reset = diff_get_color(color_diff, DIFF_RESET);
527 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
528 const char *a_prefix, *b_prefix;
529 char *data_one, *data_two;
530 size_t size_one, size_two;
531 struct emit_callback ecbdata;
532 char *line_prefix = "";
533 struct strbuf *msgbuf;
534
535 if (o && o->output_prefix) {
536 msgbuf = o->output_prefix(o, o->output_prefix_data);
537 line_prefix = msgbuf->buf;
538 }
539
540 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
541 a_prefix = o->b_prefix;
542 b_prefix = o->a_prefix;
543 } else {
544 a_prefix = o->a_prefix;
545 b_prefix = o->b_prefix;
546 }
547
548 name_a += (*name_a == '/');
549 name_b += (*name_b == '/');
550 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
551 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
552
553 strbuf_reset(&a_name);
554 strbuf_reset(&b_name);
555 quote_two_c_style(&a_name, a_prefix, name_a, 0);
556 quote_two_c_style(&b_name, b_prefix, name_b, 0);
557
558 size_one = fill_textconv(textconv_one, one, &data_one);
559 size_two = fill_textconv(textconv_two, two, &data_two);
560
561 memset(&ecbdata, 0, sizeof(ecbdata));
562 ecbdata.color_diff = color_diff;
563 ecbdata.found_changesp = &o->found_changes;
564 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
565 ecbdata.opt = o;
566 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
567 mmfile_t mf1, mf2;
568 mf1.ptr = (char *)data_one;
569 mf2.ptr = (char *)data_two;
570 mf1.size = size_one;
571 mf2.size = size_two;
572 check_blank_at_eof(&mf1, &mf2, &ecbdata);
573 }
574 ecbdata.lno_in_preimage = 1;
575 ecbdata.lno_in_postimage = 1;
576
577 lc_a = count_lines(data_one, size_one);
578 lc_b = count_lines(data_two, size_two);
579 fprintf(o->file,
580 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
581 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
582 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
583 line_prefix, fraginfo);
584 if (!o->irreversible_delete)
585 print_line_count(o->file, lc_a);
586 else
587 fprintf(o->file, "?,?");
588 fprintf(o->file, " +");
589 print_line_count(o->file, lc_b);
590 fprintf(o->file, " @@%s\n", reset);
591 if (lc_a && !o->irreversible_delete)
592 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
593 if (lc_b)
594 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
595 if (textconv_one)
596 free((char *)data_one);
597 if (textconv_two)
598 free((char *)data_two);
599 }
600
601 struct diff_words_buffer {
602 mmfile_t text;
603 long alloc;
604 struct diff_words_orig {
605 const char *begin, *end;
606 } *orig;
607 int orig_nr, orig_alloc;
608 };
609
610 static void diff_words_append(char *line, unsigned long len,
611 struct diff_words_buffer *buffer)
612 {
613 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
614 line++;
615 len--;
616 memcpy(buffer->text.ptr + buffer->text.size, line, len);
617 buffer->text.size += len;
618 buffer->text.ptr[buffer->text.size] = '\0';
619 }
620
621 struct diff_words_style_elem {
622 const char *prefix;
623 const char *suffix;
624 const char *color; /* NULL; filled in by the setup code if
625 * color is enabled */
626 };
627
628 struct diff_words_style {
629 enum diff_words_type type;
630 struct diff_words_style_elem new, old, ctx;
631 const char *newline;
632 };
633
634 static struct diff_words_style diff_words_styles[] = {
635 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
636 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
637 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
638 };
639
640 struct diff_words_data {
641 struct diff_words_buffer minus, plus;
642 const char *current_plus;
643 int last_minus;
644 struct diff_options *opt;
645 regex_t *word_regex;
646 enum diff_words_type type;
647 struct diff_words_style *style;
648 };
649
650 static int fn_out_diff_words_write_helper(FILE *fp,
651 struct diff_words_style_elem *st_el,
652 const char *newline,
653 size_t count, const char *buf,
654 const char *line_prefix)
655 {
656 int print = 0;
657
658 while (count) {
659 char *p = memchr(buf, '\n', count);
660 if (print)
661 fputs(line_prefix, fp);
662 if (p != buf) {
663 if (st_el->color && fputs(st_el->color, fp) < 0)
664 return -1;
665 if (fputs(st_el->prefix, fp) < 0 ||
666 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
667 fputs(st_el->suffix, fp) < 0)
668 return -1;
669 if (st_el->color && *st_el->color
670 && fputs(GIT_COLOR_RESET, fp) < 0)
671 return -1;
672 }
673 if (!p)
674 return 0;
675 if (fputs(newline, fp) < 0)
676 return -1;
677 count -= p + 1 - buf;
678 buf = p + 1;
679 print = 1;
680 }
681 return 0;
682 }
683
684 /*
685 * '--color-words' algorithm can be described as:
686 *
687 * 1. collect a the minus/plus lines of a diff hunk, divided into
688 * minus-lines and plus-lines;
689 *
690 * 2. break both minus-lines and plus-lines into words and
691 * place them into two mmfile_t with one word for each line;
692 *
693 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
694 *
695 * And for the common parts of the both file, we output the plus side text.
696 * diff_words->current_plus is used to trace the current position of the plus file
697 * which printed. diff_words->last_minus is used to trace the last minus word
698 * printed.
699 *
700 * For '--graph' to work with '--color-words', we need to output the graph prefix
701 * on each line of color words output. Generally, there are two conditions on
702 * which we should output the prefix.
703 *
704 * 1. diff_words->last_minus == 0 &&
705 * diff_words->current_plus == diff_words->plus.text.ptr
706 *
707 * that is: the plus text must start as a new line, and if there is no minus
708 * word printed, a graph prefix must be printed.
709 *
710 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
711 * *(diff_words->current_plus - 1) == '\n'
712 *
713 * that is: a graph prefix must be printed following a '\n'
714 */
715 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
716 {
717 if ((diff_words->last_minus == 0 &&
718 diff_words->current_plus == diff_words->plus.text.ptr) ||
719 (diff_words->current_plus > diff_words->plus.text.ptr &&
720 *(diff_words->current_plus - 1) == '\n')) {
721 return 1;
722 } else {
723 return 0;
724 }
725 }
726
727 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
728 {
729 struct diff_words_data *diff_words = priv;
730 struct diff_words_style *style = diff_words->style;
731 int minus_first, minus_len, plus_first, plus_len;
732 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
733 struct diff_options *opt = diff_words->opt;
734 struct strbuf *msgbuf;
735 char *line_prefix = "";
736
737 if (line[0] != '@' || parse_hunk_header(line, len,
738 &minus_first, &minus_len, &plus_first, &plus_len))
739 return;
740
741 assert(opt);
742 if (opt->output_prefix) {
743 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
744 line_prefix = msgbuf->buf;
745 }
746
747 /* POSIX requires that first be decremented by one if len == 0... */
748 if (minus_len) {
749 minus_begin = diff_words->minus.orig[minus_first].begin;
750 minus_end =
751 diff_words->minus.orig[minus_first + minus_len - 1].end;
752 } else
753 minus_begin = minus_end =
754 diff_words->minus.orig[minus_first].end;
755
756 if (plus_len) {
757 plus_begin = diff_words->plus.orig[plus_first].begin;
758 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
759 } else
760 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
761
762 if (color_words_output_graph_prefix(diff_words)) {
763 fputs(line_prefix, diff_words->opt->file);
764 }
765 if (diff_words->current_plus != plus_begin) {
766 fn_out_diff_words_write_helper(diff_words->opt->file,
767 &style->ctx, style->newline,
768 plus_begin - diff_words->current_plus,
769 diff_words->current_plus, line_prefix);
770 if (*(plus_begin - 1) == '\n')
771 fputs(line_prefix, diff_words->opt->file);
772 }
773 if (minus_begin != minus_end) {
774 fn_out_diff_words_write_helper(diff_words->opt->file,
775 &style->old, style->newline,
776 minus_end - minus_begin, minus_begin,
777 line_prefix);
778 }
779 if (plus_begin != plus_end) {
780 fn_out_diff_words_write_helper(diff_words->opt->file,
781 &style->new, style->newline,
782 plus_end - plus_begin, plus_begin,
783 line_prefix);
784 }
785
786 diff_words->current_plus = plus_end;
787 diff_words->last_minus = minus_first;
788 }
789
790 /* This function starts looking at *begin, and returns 0 iff a word was found. */
791 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
792 int *begin, int *end)
793 {
794 if (word_regex && *begin < buffer->size) {
795 regmatch_t match[1];
796 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
797 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
798 '\n', match[0].rm_eo - match[0].rm_so);
799 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
800 *begin += match[0].rm_so;
801 return *begin >= *end;
802 }
803 return -1;
804 }
805
806 /* find the next word */
807 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
808 (*begin)++;
809 if (*begin >= buffer->size)
810 return -1;
811
812 /* find the end of the word */
813 *end = *begin + 1;
814 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
815 (*end)++;
816
817 return 0;
818 }
819
820 /*
821 * This function splits the words in buffer->text, stores the list with
822 * newline separator into out, and saves the offsets of the original words
823 * in buffer->orig.
824 */
825 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
826 regex_t *word_regex)
827 {
828 int i, j;
829 long alloc = 0;
830
831 out->size = 0;
832 out->ptr = NULL;
833
834 /* fake an empty "0th" word */
835 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
836 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
837 buffer->orig_nr = 1;
838
839 for (i = 0; i < buffer->text.size; i++) {
840 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
841 return;
842
843 /* store original boundaries */
844 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
845 buffer->orig_alloc);
846 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
847 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
848 buffer->orig_nr++;
849
850 /* store one word */
851 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
852 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
853 out->ptr[out->size + j - i] = '\n';
854 out->size += j - i + 1;
855
856 i = j - 1;
857 }
858 }
859
860 /* this executes the word diff on the accumulated buffers */
861 static void diff_words_show(struct diff_words_data *diff_words)
862 {
863 xpparam_t xpp;
864 xdemitconf_t xecfg;
865 mmfile_t minus, plus;
866 struct diff_words_style *style = diff_words->style;
867
868 struct diff_options *opt = diff_words->opt;
869 struct strbuf *msgbuf;
870 char *line_prefix = "";
871
872 assert(opt);
873 if (opt->output_prefix) {
874 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
875 line_prefix = msgbuf->buf;
876 }
877
878 /* special case: only removal */
879 if (!diff_words->plus.text.size) {
880 fputs(line_prefix, diff_words->opt->file);
881 fn_out_diff_words_write_helper(diff_words->opt->file,
882 &style->old, style->newline,
883 diff_words->minus.text.size,
884 diff_words->minus.text.ptr, line_prefix);
885 diff_words->minus.text.size = 0;
886 return;
887 }
888
889 diff_words->current_plus = diff_words->plus.text.ptr;
890 diff_words->last_minus = 0;
891
892 memset(&xpp, 0, sizeof(xpp));
893 memset(&xecfg, 0, sizeof(xecfg));
894 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
895 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
896 xpp.flags = 0;
897 /* as only the hunk header will be parsed, we need a 0-context */
898 xecfg.ctxlen = 0;
899 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
900 &xpp, &xecfg);
901 free(minus.ptr);
902 free(plus.ptr);
903 if (diff_words->current_plus != diff_words->plus.text.ptr +
904 diff_words->plus.text.size) {
905 if (color_words_output_graph_prefix(diff_words))
906 fputs(line_prefix, diff_words->opt->file);
907 fn_out_diff_words_write_helper(diff_words->opt->file,
908 &style->ctx, style->newline,
909 diff_words->plus.text.ptr + diff_words->plus.text.size
910 - diff_words->current_plus, diff_words->current_plus,
911 line_prefix);
912 }
913 diff_words->minus.text.size = diff_words->plus.text.size = 0;
914 }
915
916 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
917 static void diff_words_flush(struct emit_callback *ecbdata)
918 {
919 if (ecbdata->diff_words->minus.text.size ||
920 ecbdata->diff_words->plus.text.size)
921 diff_words_show(ecbdata->diff_words);
922 }
923
924 static void free_diff_words_data(struct emit_callback *ecbdata)
925 {
926 if (ecbdata->diff_words) {
927 diff_words_flush(ecbdata);
928 free (ecbdata->diff_words->minus.text.ptr);
929 free (ecbdata->diff_words->minus.orig);
930 free (ecbdata->diff_words->plus.text.ptr);
931 free (ecbdata->diff_words->plus.orig);
932 if (ecbdata->diff_words->word_regex) {
933 regfree(ecbdata->diff_words->word_regex);
934 free(ecbdata->diff_words->word_regex);
935 }
936 free(ecbdata->diff_words);
937 ecbdata->diff_words = NULL;
938 }
939 }
940
941 const char *diff_get_color(int diff_use_color, enum color_diff ix)
942 {
943 if (diff_use_color)
944 return diff_colors[ix];
945 return "";
946 }
947
948 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
949 {
950 const char *cp;
951 unsigned long allot;
952 size_t l = len;
953
954 if (ecb->truncate)
955 return ecb->truncate(line, len);
956 cp = line;
957 allot = l;
958 while (0 < l) {
959 (void) utf8_width(&cp, &l);
960 if (!cp)
961 break; /* truncated in the middle? */
962 }
963 return allot - l;
964 }
965
966 static void find_lno(const char *line, struct emit_callback *ecbdata)
967 {
968 const char *p;
969 ecbdata->lno_in_preimage = 0;
970 ecbdata->lno_in_postimage = 0;
971 p = strchr(line, '-');
972 if (!p)
973 return; /* cannot happen */
974 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
975 p = strchr(p, '+');
976 if (!p)
977 return; /* cannot happen */
978 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
979 }
980
981 static void fn_out_consume(void *priv, char *line, unsigned long len)
982 {
983 struct emit_callback *ecbdata = priv;
984 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
985 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
986 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
987 struct diff_options *o = ecbdata->opt;
988 char *line_prefix = "";
989 struct strbuf *msgbuf;
990
991 if (o && o->output_prefix) {
992 msgbuf = o->output_prefix(o, o->output_prefix_data);
993 line_prefix = msgbuf->buf;
994 }
995
996 if (ecbdata->header) {
997 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
998 strbuf_reset(ecbdata->header);
999 ecbdata->header = NULL;
1000 }
1001 *(ecbdata->found_changesp) = 1;
1002
1003 if (ecbdata->label_path[0]) {
1004 const char *name_a_tab, *name_b_tab;
1005
1006 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1007 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1008
1009 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1010 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1011 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1012 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1013 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1014 }
1015
1016 if (diff_suppress_blank_empty
1017 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1018 line[0] = '\n';
1019 len = 1;
1020 }
1021
1022 if (line[0] == '@') {
1023 if (ecbdata->diff_words)
1024 diff_words_flush(ecbdata);
1025 len = sane_truncate_line(ecbdata, line, len);
1026 find_lno(line, ecbdata);
1027 emit_hunk_header(ecbdata, line, len);
1028 if (line[len-1] != '\n')
1029 putc('\n', ecbdata->opt->file);
1030 return;
1031 }
1032
1033 if (len < 1) {
1034 emit_line(ecbdata->opt, reset, reset, line, len);
1035 if (ecbdata->diff_words
1036 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1037 fputs("~\n", ecbdata->opt->file);
1038 return;
1039 }
1040
1041 if (ecbdata->diff_words) {
1042 if (line[0] == '-') {
1043 diff_words_append(line, len,
1044 &ecbdata->diff_words->minus);
1045 return;
1046 } else if (line[0] == '+') {
1047 diff_words_append(line, len,
1048 &ecbdata->diff_words->plus);
1049 return;
1050 }
1051 diff_words_flush(ecbdata);
1052 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1053 emit_line(ecbdata->opt, plain, reset, line, len);
1054 fputs("~\n", ecbdata->opt->file);
1055 } else {
1056 /* don't print the prefix character */
1057 emit_line(ecbdata->opt, plain, reset, line+1, len-1);
1058 }
1059 return;
1060 }
1061
1062 if (line[0] != '+') {
1063 const char *color =
1064 diff_get_color(ecbdata->color_diff,
1065 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1066 ecbdata->lno_in_preimage++;
1067 if (line[0] == ' ')
1068 ecbdata->lno_in_postimage++;
1069 emit_line(ecbdata->opt, color, reset, line, len);
1070 } else {
1071 ecbdata->lno_in_postimage++;
1072 emit_add_line(reset, ecbdata, line + 1, len - 1);
1073 }
1074 }
1075
1076 static char *pprint_rename(const char *a, const char *b)
1077 {
1078 const char *old = a;
1079 const char *new = b;
1080 struct strbuf name = STRBUF_INIT;
1081 int pfx_length, sfx_length;
1082 int len_a = strlen(a);
1083 int len_b = strlen(b);
1084 int a_midlen, b_midlen;
1085 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1086 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1087
1088 if (qlen_a || qlen_b) {
1089 quote_c_style(a, &name, NULL, 0);
1090 strbuf_addstr(&name, " => ");
1091 quote_c_style(b, &name, NULL, 0);
1092 return strbuf_detach(&name, NULL);
1093 }
1094
1095 /* Find common prefix */
1096 pfx_length = 0;
1097 while (*old && *new && *old == *new) {
1098 if (*old == '/')
1099 pfx_length = old - a + 1;
1100 old++;
1101 new++;
1102 }
1103
1104 /* Find common suffix */
1105 old = a + len_a;
1106 new = b + len_b;
1107 sfx_length = 0;
1108 while (a <= old && b <= new && *old == *new) {
1109 if (*old == '/')
1110 sfx_length = len_a - (old - a);
1111 old--;
1112 new--;
1113 }
1114
1115 /*
1116 * pfx{mid-a => mid-b}sfx
1117 * {pfx-a => pfx-b}sfx
1118 * pfx{sfx-a => sfx-b}
1119 * name-a => name-b
1120 */
1121 a_midlen = len_a - pfx_length - sfx_length;
1122 b_midlen = len_b - pfx_length - sfx_length;
1123 if (a_midlen < 0)
1124 a_midlen = 0;
1125 if (b_midlen < 0)
1126 b_midlen = 0;
1127
1128 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1129 if (pfx_length + sfx_length) {
1130 strbuf_add(&name, a, pfx_length);
1131 strbuf_addch(&name, '{');
1132 }
1133 strbuf_add(&name, a + pfx_length, a_midlen);
1134 strbuf_addstr(&name, " => ");
1135 strbuf_add(&name, b + pfx_length, b_midlen);
1136 if (pfx_length + sfx_length) {
1137 strbuf_addch(&name, '}');
1138 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1139 }
1140 return strbuf_detach(&name, NULL);
1141 }
1142
1143 struct diffstat_t {
1144 int nr;
1145 int alloc;
1146 struct diffstat_file {
1147 char *from_name;
1148 char *name;
1149 char *print_name;
1150 unsigned is_unmerged:1;
1151 unsigned is_binary:1;
1152 unsigned is_renamed:1;
1153 uintmax_t added, deleted;
1154 } **files;
1155 };
1156
1157 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1158 const char *name_a,
1159 const char *name_b)
1160 {
1161 struct diffstat_file *x;
1162 x = xcalloc(sizeof (*x), 1);
1163 if (diffstat->nr == diffstat->alloc) {
1164 diffstat->alloc = alloc_nr(diffstat->alloc);
1165 diffstat->files = xrealloc(diffstat->files,
1166 diffstat->alloc * sizeof(x));
1167 }
1168 diffstat->files[diffstat->nr++] = x;
1169 if (name_b) {
1170 x->from_name = xstrdup(name_a);
1171 x->name = xstrdup(name_b);
1172 x->is_renamed = 1;
1173 }
1174 else {
1175 x->from_name = NULL;
1176 x->name = xstrdup(name_a);
1177 }
1178 return x;
1179 }
1180
1181 static void diffstat_consume(void *priv, char *line, unsigned long len)
1182 {
1183 struct diffstat_t *diffstat = priv;
1184 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1185
1186 if (line[0] == '+')
1187 x->added++;
1188 else if (line[0] == '-')
1189 x->deleted++;
1190 }
1191
1192 const char mime_boundary_leader[] = "------------";
1193
1194 static int scale_linear(int it, int width, int max_change)
1195 {
1196 /*
1197 * make sure that at least one '-' is printed if there were deletions,
1198 * and likewise for '+'.
1199 */
1200 if (max_change < 2)
1201 return it;
1202 return ((it - 1) * (width - 1) + max_change - 1) / (max_change - 1);
1203 }
1204
1205 static void show_name(FILE *file,
1206 const char *prefix, const char *name, int len)
1207 {
1208 fprintf(file, " %s%-*s |", prefix, len, name);
1209 }
1210
1211 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1212 {
1213 if (cnt <= 0)
1214 return;
1215 fprintf(file, "%s", set);
1216 while (cnt--)
1217 putc(ch, file);
1218 fprintf(file, "%s", reset);
1219 }
1220
1221 static void fill_print_name(struct diffstat_file *file)
1222 {
1223 char *pname;
1224
1225 if (file->print_name)
1226 return;
1227
1228 if (!file->is_renamed) {
1229 struct strbuf buf = STRBUF_INIT;
1230 if (quote_c_style(file->name, &buf, NULL, 0)) {
1231 pname = strbuf_detach(&buf, NULL);
1232 } else {
1233 pname = file->name;
1234 strbuf_release(&buf);
1235 }
1236 } else {
1237 pname = pprint_rename(file->from_name, file->name);
1238 }
1239 file->print_name = pname;
1240 }
1241
1242 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1243 {
1244 int i, len, add, del, adds = 0, dels = 0;
1245 uintmax_t max_change = 0, max_len = 0;
1246 int total_files = data->nr;
1247 int width, name_width, count;
1248 const char *reset, *add_c, *del_c;
1249 const char *line_prefix = "";
1250 int extra_shown = 0;
1251 struct strbuf *msg = NULL;
1252
1253 if (data->nr == 0)
1254 return;
1255
1256 if (options->output_prefix) {
1257 msg = options->output_prefix(options, options->output_prefix_data);
1258 line_prefix = msg->buf;
1259 }
1260
1261 width = options->stat_width ? options->stat_width : 80;
1262 name_width = options->stat_name_width ? options->stat_name_width : 50;
1263 count = options->stat_count ? options->stat_count : data->nr;
1264
1265 /* Sanity: give at least 5 columns to the graph,
1266 * but leave at least 10 columns for the name.
1267 */
1268 if (width < 25)
1269 width = 25;
1270 if (name_width < 10)
1271 name_width = 10;
1272 else if (width < name_width + 15)
1273 name_width = width - 15;
1274
1275 /* Find the longest filename and max number of changes */
1276 reset = diff_get_color_opt(options, DIFF_RESET);
1277 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1278 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1279
1280 for (i = 0; (i < count) && (i < data->nr); i++) {
1281 struct diffstat_file *file = data->files[i];
1282 uintmax_t change = file->added + file->deleted;
1283 if (!data->files[i]->is_renamed &&
1284 (change == 0)) {
1285 count++; /* not shown == room for one more */
1286 continue;
1287 }
1288 fill_print_name(file);
1289 len = strlen(file->print_name);
1290 if (max_len < len)
1291 max_len = len;
1292
1293 if (file->is_binary || file->is_unmerged)
1294 continue;
1295 if (max_change < change)
1296 max_change = change;
1297 }
1298 count = i; /* min(count, data->nr) */
1299
1300 /* Compute the width of the graph part;
1301 * 10 is for one blank at the beginning of the line plus
1302 * " | count " between the name and the graph.
1303 *
1304 * From here on, name_width is the width of the name area,
1305 * and width is the width of the graph area.
1306 */
1307 name_width = (name_width < max_len) ? name_width : max_len;
1308 if (width < (name_width + 10) + max_change)
1309 width = width - (name_width + 10);
1310 else
1311 width = max_change;
1312
1313 for (i = 0; i < count; i++) {
1314 const char *prefix = "";
1315 char *name = data->files[i]->print_name;
1316 uintmax_t added = data->files[i]->added;
1317 uintmax_t deleted = data->files[i]->deleted;
1318 int name_len;
1319
1320 if (!data->files[i]->is_renamed &&
1321 (added + deleted == 0)) {
1322 total_files--;
1323 continue;
1324 }
1325 /*
1326 * "scale" the filename
1327 */
1328 len = name_width;
1329 name_len = strlen(name);
1330 if (name_width < name_len) {
1331 char *slash;
1332 prefix = "...";
1333 len -= 3;
1334 name += name_len - len;
1335 slash = strchr(name, '/');
1336 if (slash)
1337 name = slash;
1338 }
1339
1340 if (data->files[i]->is_binary) {
1341 fprintf(options->file, "%s", line_prefix);
1342 show_name(options->file, prefix, name, len);
1343 fprintf(options->file, " Bin ");
1344 fprintf(options->file, "%s%"PRIuMAX"%s",
1345 del_c, deleted, reset);
1346 fprintf(options->file, " -> ");
1347 fprintf(options->file, "%s%"PRIuMAX"%s",
1348 add_c, added, reset);
1349 fprintf(options->file, " bytes");
1350 fprintf(options->file, "\n");
1351 continue;
1352 }
1353 else if (data->files[i]->is_unmerged) {
1354 fprintf(options->file, "%s", line_prefix);
1355 show_name(options->file, prefix, name, len);
1356 fprintf(options->file, " Unmerged\n");
1357 continue;
1358 }
1359
1360 /*
1361 * scale the add/delete
1362 */
1363 add = added;
1364 del = deleted;
1365 adds += add;
1366 dels += del;
1367
1368 if (width <= max_change) {
1369 add = scale_linear(add, width, max_change);
1370 del = scale_linear(del, width, max_change);
1371 }
1372 fprintf(options->file, "%s", line_prefix);
1373 show_name(options->file, prefix, name, len);
1374 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1375 added + deleted ? " " : "");
1376 show_graph(options->file, '+', add, add_c, reset);
1377 show_graph(options->file, '-', del, del_c, reset);
1378 fprintf(options->file, "\n");
1379 }
1380 for (i = count; i < data->nr; i++) {
1381 uintmax_t added = data->files[i]->added;
1382 uintmax_t deleted = data->files[i]->deleted;
1383 if (!data->files[i]->is_renamed &&
1384 (added + deleted == 0)) {
1385 total_files--;
1386 continue;
1387 }
1388 adds += added;
1389 dels += deleted;
1390 if (!extra_shown)
1391 fprintf(options->file, "%s ...\n", line_prefix);
1392 extra_shown = 1;
1393 }
1394 fprintf(options->file, "%s", line_prefix);
1395 fprintf(options->file,
1396 " %d files changed, %d insertions(+), %d deletions(-)\n",
1397 total_files, adds, dels);
1398 }
1399
1400 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1401 {
1402 int i, adds = 0, dels = 0, total_files = data->nr;
1403
1404 if (data->nr == 0)
1405 return;
1406
1407 for (i = 0; i < data->nr; i++) {
1408 if (!data->files[i]->is_binary &&
1409 !data->files[i]->is_unmerged) {
1410 int added = data->files[i]->added;
1411 int deleted= data->files[i]->deleted;
1412 if (!data->files[i]->is_renamed &&
1413 (added + deleted == 0)) {
1414 total_files--;
1415 } else {
1416 adds += added;
1417 dels += deleted;
1418 }
1419 }
1420 }
1421 if (options->output_prefix) {
1422 struct strbuf *msg = NULL;
1423 msg = options->output_prefix(options,
1424 options->output_prefix_data);
1425 fprintf(options->file, "%s", msg->buf);
1426 }
1427 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1428 total_files, adds, dels);
1429 }
1430
1431 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1432 {
1433 int i;
1434
1435 if (data->nr == 0)
1436 return;
1437
1438 for (i = 0; i < data->nr; i++) {
1439 struct diffstat_file *file = data->files[i];
1440
1441 if (options->output_prefix) {
1442 struct strbuf *msg = NULL;
1443 msg = options->output_prefix(options,
1444 options->output_prefix_data);
1445 fprintf(options->file, "%s", msg->buf);
1446 }
1447
1448 if (file->is_binary)
1449 fprintf(options->file, "-\t-\t");
1450 else
1451 fprintf(options->file,
1452 "%"PRIuMAX"\t%"PRIuMAX"\t",
1453 file->added, file->deleted);
1454 if (options->line_termination) {
1455 fill_print_name(file);
1456 if (!file->is_renamed)
1457 write_name_quoted(file->name, options->file,
1458 options->line_termination);
1459 else {
1460 fputs(file->print_name, options->file);
1461 putc(options->line_termination, options->file);
1462 }
1463 } else {
1464 if (file->is_renamed) {
1465 putc('\0', options->file);
1466 write_name_quoted(file->from_name, options->file, '\0');
1467 }
1468 write_name_quoted(file->name, options->file, '\0');
1469 }
1470 }
1471 }
1472
1473 struct dirstat_file {
1474 const char *name;
1475 unsigned long changed;
1476 };
1477
1478 struct dirstat_dir {
1479 struct dirstat_file *files;
1480 int alloc, nr, percent, cumulative;
1481 };
1482
1483 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1484 unsigned long changed, const char *base, int baselen)
1485 {
1486 unsigned long this_dir = 0;
1487 unsigned int sources = 0;
1488 const char *line_prefix = "";
1489 struct strbuf *msg = NULL;
1490
1491 if (opt->output_prefix) {
1492 msg = opt->output_prefix(opt, opt->output_prefix_data);
1493 line_prefix = msg->buf;
1494 }
1495
1496 while (dir->nr) {
1497 struct dirstat_file *f = dir->files;
1498 int namelen = strlen(f->name);
1499 unsigned long this;
1500 char *slash;
1501
1502 if (namelen < baselen)
1503 break;
1504 if (memcmp(f->name, base, baselen))
1505 break;
1506 slash = strchr(f->name + baselen, '/');
1507 if (slash) {
1508 int newbaselen = slash + 1 - f->name;
1509 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1510 sources++;
1511 } else {
1512 this = f->changed;
1513 dir->files++;
1514 dir->nr--;
1515 sources += 2;
1516 }
1517 this_dir += this;
1518 }
1519
1520 /*
1521 * We don't report dirstat's for
1522 * - the top level
1523 * - or cases where everything came from a single directory
1524 * under this directory (sources == 1).
1525 */
1526 if (baselen && sources != 1) {
1527 int permille = this_dir * 1000 / changed;
1528 if (permille) {
1529 int percent = permille / 10;
1530 if (percent >= dir->percent) {
1531 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1532 percent, permille % 10, baselen, base);
1533 if (!dir->cumulative)
1534 return 0;
1535 }
1536 }
1537 }
1538 return this_dir;
1539 }
1540
1541 static int dirstat_compare(const void *_a, const void *_b)
1542 {
1543 const struct dirstat_file *a = _a;
1544 const struct dirstat_file *b = _b;
1545 return strcmp(a->name, b->name);
1546 }
1547
1548 static void show_dirstat(struct diff_options *options)
1549 {
1550 int i;
1551 unsigned long changed;
1552 struct dirstat_dir dir;
1553 struct diff_queue_struct *q = &diff_queued_diff;
1554
1555 dir.files = NULL;
1556 dir.alloc = 0;
1557 dir.nr = 0;
1558 dir.percent = options->dirstat_percent;
1559 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1560
1561 changed = 0;
1562 for (i = 0; i < q->nr; i++) {
1563 struct diff_filepair *p = q->queue[i];
1564 const char *name;
1565 unsigned long copied, added, damage;
1566 int content_changed;
1567
1568 name = p->two->path ? p->two->path : p->one->path;
1569
1570 if (p->one->sha1_valid && p->two->sha1_valid)
1571 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1572 else
1573 content_changed = 1;
1574
1575 if (!content_changed) {
1576 /*
1577 * The SHA1 has not changed, so pre-/post-content is
1578 * identical. We can therefore skip looking at the
1579 * file contents altogether.
1580 */
1581 damage = 0;
1582 goto found_damage;
1583 }
1584
1585 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1586 /*
1587 * In --dirstat-by-file mode, we don't really need to
1588 * look at the actual file contents at all.
1589 * The fact that the SHA1 changed is enough for us to
1590 * add this file to the list of results
1591 * (with each file contributing equal damage).
1592 */
1593 damage = 1;
1594 goto found_damage;
1595 }
1596
1597 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1598 diff_populate_filespec(p->one, 0);
1599 diff_populate_filespec(p->two, 0);
1600 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1601 &copied, &added);
1602 diff_free_filespec_data(p->one);
1603 diff_free_filespec_data(p->two);
1604 } else if (DIFF_FILE_VALID(p->one)) {
1605 diff_populate_filespec(p->one, 1);
1606 copied = added = 0;
1607 diff_free_filespec_data(p->one);
1608 } else if (DIFF_FILE_VALID(p->two)) {
1609 diff_populate_filespec(p->two, 1);
1610 copied = 0;
1611 added = p->two->size;
1612 diff_free_filespec_data(p->two);
1613 } else
1614 continue;
1615
1616 /*
1617 * Original minus copied is the removed material,
1618 * added is the new material. They are both damages
1619 * made to the preimage.
1620 * If the resulting damage is zero, we know that
1621 * diffcore_count_changes() considers the two entries to
1622 * be identical, but since content_changed is true, we
1623 * know that there must have been _some_ kind of change,
1624 * so we force all entries to have damage > 0.
1625 */
1626 damage = (p->one->size - copied) + added;
1627 if (!damage)
1628 damage = 1;
1629
1630 found_damage:
1631 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1632 dir.files[dir.nr].name = name;
1633 dir.files[dir.nr].changed = damage;
1634 changed += damage;
1635 dir.nr++;
1636 }
1637
1638 /* This can happen even with many files, if everything was renames */
1639 if (!changed)
1640 return;
1641
1642 /* Show all directories with more than x% of the changes */
1643 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1644 gather_dirstat(options, &dir, changed, "", 0);
1645 }
1646
1647 static void free_diffstat_info(struct diffstat_t *diffstat)
1648 {
1649 int i;
1650 for (i = 0; i < diffstat->nr; i++) {
1651 struct diffstat_file *f = diffstat->files[i];
1652 if (f->name != f->print_name)
1653 free(f->print_name);
1654 free(f->name);
1655 free(f->from_name);
1656 free(f);
1657 }
1658 free(diffstat->files);
1659 }
1660
1661 struct checkdiff_t {
1662 const char *filename;
1663 int lineno;
1664 int conflict_marker_size;
1665 struct diff_options *o;
1666 unsigned ws_rule;
1667 unsigned status;
1668 };
1669
1670 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1671 {
1672 char firstchar;
1673 int cnt;
1674
1675 if (len < marker_size + 1)
1676 return 0;
1677 firstchar = line[0];
1678 switch (firstchar) {
1679 case '=': case '>': case '<': case '|':
1680 break;
1681 default:
1682 return 0;
1683 }
1684 for (cnt = 1; cnt < marker_size; cnt++)
1685 if (line[cnt] != firstchar)
1686 return 0;
1687 /* line[1] thru line[marker_size-1] are same as firstchar */
1688 if (len < marker_size + 1 || !isspace(line[marker_size]))
1689 return 0;
1690 return 1;
1691 }
1692
1693 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1694 {
1695 struct checkdiff_t *data = priv;
1696 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1697 int marker_size = data->conflict_marker_size;
1698 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1699 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1700 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1701 char *err;
1702 char *line_prefix = "";
1703 struct strbuf *msgbuf;
1704
1705 assert(data->o);
1706 if (data->o->output_prefix) {
1707 msgbuf = data->o->output_prefix(data->o,
1708 data->o->output_prefix_data);
1709 line_prefix = msgbuf->buf;
1710 }
1711
1712 if (line[0] == '+') {
1713 unsigned bad;
1714 data->lineno++;
1715 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1716 data->status |= 1;
1717 fprintf(data->o->file,
1718 "%s%s:%d: leftover conflict marker\n",
1719 line_prefix, data->filename, data->lineno);
1720 }
1721 bad = ws_check(line + 1, len - 1, data->ws_rule);
1722 if (!bad)
1723 return;
1724 data->status |= bad;
1725 err = whitespace_error_string(bad);
1726 fprintf(data->o->file, "%s%s:%d: %s.\n",
1727 line_prefix, data->filename, data->lineno, err);
1728 free(err);
1729 emit_line(data->o, set, reset, line, 1);
1730 ws_check_emit(line + 1, len - 1, data->ws_rule,
1731 data->o->file, set, reset, ws);
1732 } else if (line[0] == ' ') {
1733 data->lineno++;
1734 } else if (line[0] == '@') {
1735 char *plus = strchr(line, '+');
1736 if (plus)
1737 data->lineno = strtol(plus, NULL, 10) - 1;
1738 else
1739 die("invalid diff");
1740 }
1741 }
1742
1743 static unsigned char *deflate_it(char *data,
1744 unsigned long size,
1745 unsigned long *result_size)
1746 {
1747 int bound;
1748 unsigned char *deflated;
1749 z_stream stream;
1750
1751 memset(&stream, 0, sizeof(stream));
1752 deflateInit(&stream, zlib_compression_level);
1753 bound = deflateBound(&stream, size);
1754 deflated = xmalloc(bound);
1755 stream.next_out = deflated;
1756 stream.avail_out = bound;
1757
1758 stream.next_in = (unsigned char *)data;
1759 stream.avail_in = size;
1760 while (deflate(&stream, Z_FINISH) == Z_OK)
1761 ; /* nothing */
1762 deflateEnd(&stream);
1763 *result_size = stream.total_out;
1764 return deflated;
1765 }
1766
1767 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1768 {
1769 void *cp;
1770 void *delta;
1771 void *deflated;
1772 void *data;
1773 unsigned long orig_size;
1774 unsigned long delta_size;
1775 unsigned long deflate_size;
1776 unsigned long data_size;
1777
1778 /* We could do deflated delta, or we could do just deflated two,
1779 * whichever is smaller.
1780 */
1781 delta = NULL;
1782 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1783 if (one->size && two->size) {
1784 delta = diff_delta(one->ptr, one->size,
1785 two->ptr, two->size,
1786 &delta_size, deflate_size);
1787 if (delta) {
1788 void *to_free = delta;
1789 orig_size = delta_size;
1790 delta = deflate_it(delta, delta_size, &delta_size);
1791 free(to_free);
1792 }
1793 }
1794
1795 if (delta && delta_size < deflate_size) {
1796 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1797 free(deflated);
1798 data = delta;
1799 data_size = delta_size;
1800 }
1801 else {
1802 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1803 free(delta);
1804 data = deflated;
1805 data_size = deflate_size;
1806 }
1807
1808 /* emit data encoded in base85 */
1809 cp = data;
1810 while (data_size) {
1811 int bytes = (52 < data_size) ? 52 : data_size;
1812 char line[70];
1813 data_size -= bytes;
1814 if (bytes <= 26)
1815 line[0] = bytes + 'A' - 1;
1816 else
1817 line[0] = bytes - 26 + 'a' - 1;
1818 encode_85(line + 1, cp, bytes);
1819 cp = (char *) cp + bytes;
1820 fprintf(file, "%s", prefix);
1821 fputs(line, file);
1822 fputc('\n', file);
1823 }
1824 fprintf(file, "%s\n", prefix);
1825 free(data);
1826 }
1827
1828 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1829 {
1830 fprintf(file, "%sGIT binary patch\n", prefix);
1831 emit_binary_diff_body(file, one, two, prefix);
1832 emit_binary_diff_body(file, two, one, prefix);
1833 }
1834
1835 static void diff_filespec_load_driver(struct diff_filespec *one)
1836 {
1837 /* Use already-loaded driver */
1838 if (one->driver)
1839 return;
1840
1841 if (S_ISREG(one->mode))
1842 one->driver = userdiff_find_by_path(one->path);
1843
1844 /* Fallback to default settings */
1845 if (!one->driver)
1846 one->driver = userdiff_find_by_name("default");
1847 }
1848
1849 int diff_filespec_is_binary(struct diff_filespec *one)
1850 {
1851 if (one->is_binary == -1) {
1852 diff_filespec_load_driver(one);
1853 if (one->driver->binary != -1)
1854 one->is_binary = one->driver->binary;
1855 else {
1856 if (!one->data && DIFF_FILE_VALID(one))
1857 diff_populate_filespec(one, 0);
1858 if (one->data)
1859 one->is_binary = buffer_is_binary(one->data,
1860 one->size);
1861 if (one->is_binary == -1)
1862 one->is_binary = 0;
1863 }
1864 }
1865 return one->is_binary;
1866 }
1867
1868 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1869 {
1870 diff_filespec_load_driver(one);
1871 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1872 }
1873
1874 static const char *userdiff_word_regex(struct diff_filespec *one)
1875 {
1876 diff_filespec_load_driver(one);
1877 return one->driver->word_regex;
1878 }
1879
1880 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1881 {
1882 if (!options->a_prefix)
1883 options->a_prefix = a;
1884 if (!options->b_prefix)
1885 options->b_prefix = b;
1886 }
1887
1888 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1889 {
1890 if (!DIFF_FILE_VALID(one))
1891 return NULL;
1892
1893 diff_filespec_load_driver(one);
1894 if (!one->driver->textconv)
1895 return NULL;
1896
1897 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1898 struct notes_cache *c = xmalloc(sizeof(*c));
1899 struct strbuf name = STRBUF_INIT;
1900
1901 strbuf_addf(&name, "textconv/%s", one->driver->name);
1902 notes_cache_init(c, name.buf, one->driver->textconv);
1903 one->driver->textconv_cache = c;
1904 }
1905
1906 return one->driver;
1907 }
1908
1909 static void builtin_diff(const char *name_a,
1910 const char *name_b,
1911 struct diff_filespec *one,
1912 struct diff_filespec *two,
1913 const char *xfrm_msg,
1914 int must_show_header,
1915 struct diff_options *o,
1916 int complete_rewrite)
1917 {
1918 mmfile_t mf1, mf2;
1919 const char *lbl[2];
1920 char *a_one, *b_two;
1921 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1922 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1923 const char *a_prefix, *b_prefix;
1924 struct userdiff_driver *textconv_one = NULL;
1925 struct userdiff_driver *textconv_two = NULL;
1926 struct strbuf header = STRBUF_INIT;
1927 struct strbuf *msgbuf;
1928 char *line_prefix = "";
1929
1930 if (o->output_prefix) {
1931 msgbuf = o->output_prefix(o, o->output_prefix_data);
1932 line_prefix = msgbuf->buf;
1933 }
1934
1935 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1936 (!one->mode || S_ISGITLINK(one->mode)) &&
1937 (!two->mode || S_ISGITLINK(two->mode))) {
1938 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1939 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1940 show_submodule_summary(o->file, one ? one->path : two->path,
1941 one->sha1, two->sha1, two->dirty_submodule,
1942 del, add, reset);
1943 return;
1944 }
1945
1946 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1947 textconv_one = get_textconv(one);
1948 textconv_two = get_textconv(two);
1949 }
1950
1951 diff_set_mnemonic_prefix(o, "a/", "b/");
1952 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1953 a_prefix = o->b_prefix;
1954 b_prefix = o->a_prefix;
1955 } else {
1956 a_prefix = o->a_prefix;
1957 b_prefix = o->b_prefix;
1958 }
1959
1960 /* Never use a non-valid filename anywhere if at all possible */
1961 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1962 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1963
1964 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1965 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1966 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1967 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1968 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1969 if (lbl[0][0] == '/') {
1970 /* /dev/null */
1971 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1972 if (xfrm_msg)
1973 strbuf_addstr(&header, xfrm_msg);
1974 must_show_header = 1;
1975 }
1976 else if (lbl[1][0] == '/') {
1977 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1978 if (xfrm_msg)
1979 strbuf_addstr(&header, xfrm_msg);
1980 must_show_header = 1;
1981 }
1982 else {
1983 if (one->mode != two->mode) {
1984 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1985 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1986 must_show_header = 1;
1987 }
1988 if (xfrm_msg)
1989 strbuf_addstr(&header, xfrm_msg);
1990
1991 /*
1992 * we do not run diff between different kind
1993 * of objects.
1994 */
1995 if ((one->mode ^ two->mode) & S_IFMT)
1996 goto free_ab_and_return;
1997 if (complete_rewrite &&
1998 (textconv_one || !diff_filespec_is_binary(one)) &&
1999 (textconv_two || !diff_filespec_is_binary(two))) {
2000 fprintf(o->file, "%s", header.buf);
2001 strbuf_reset(&header);
2002 emit_rewrite_diff(name_a, name_b, one, two,
2003 textconv_one, textconv_two, o);
2004 o->found_changes = 1;
2005 goto free_ab_and_return;
2006 }
2007 }
2008
2009 if (o->irreversible_delete && lbl[1][0] == '/') {
2010 fprintf(o->file, "%s", header.buf);
2011 strbuf_reset(&header);
2012 goto free_ab_and_return;
2013 } else if (!DIFF_OPT_TST(o, TEXT) &&
2014 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2015 (!textconv_two && diff_filespec_is_binary(two)) )) {
2016 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2017 die("unable to read files to diff");
2018 /* Quite common confusing case */
2019 if (mf1.size == mf2.size &&
2020 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2021 if (must_show_header)
2022 fprintf(o->file, "%s", header.buf);
2023 goto free_ab_and_return;
2024 }
2025 fprintf(o->file, "%s", header.buf);
2026 strbuf_reset(&header);
2027 if (DIFF_OPT_TST(o, BINARY))
2028 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2029 else
2030 fprintf(o->file, "%sBinary files %s and %s differ\n",
2031 line_prefix, lbl[0], lbl[1]);
2032 o->found_changes = 1;
2033 } else {
2034 /* Crazy xdl interfaces.. */
2035 const char *diffopts = getenv("GIT_DIFF_OPTS");
2036 xpparam_t xpp;
2037 xdemitconf_t xecfg;
2038 struct emit_callback ecbdata;
2039 const struct userdiff_funcname *pe;
2040
2041 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2042 fprintf(o->file, "%s", header.buf);
2043 strbuf_reset(&header);
2044 }
2045
2046 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2047 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2048
2049 pe = diff_funcname_pattern(one);
2050 if (!pe)
2051 pe = diff_funcname_pattern(two);
2052
2053 memset(&xpp, 0, sizeof(xpp));
2054 memset(&xecfg, 0, sizeof(xecfg));
2055 memset(&ecbdata, 0, sizeof(ecbdata));
2056 ecbdata.label_path = lbl;
2057 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2058 ecbdata.found_changesp = &o->found_changes;
2059 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2060 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2061 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2062 ecbdata.opt = o;
2063 ecbdata.header = header.len ? &header : NULL;
2064 xpp.flags = o->xdl_opts;
2065 xecfg.ctxlen = o->context;
2066 xecfg.interhunkctxlen = o->interhunkcontext;
2067 xecfg.flags = XDL_EMIT_FUNCNAMES;
2068 if (pe)
2069 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2070 if (!diffopts)
2071 ;
2072 else if (!prefixcmp(diffopts, "--unified="))
2073 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2074 else if (!prefixcmp(diffopts, "-u"))
2075 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2076 if (o->word_diff) {
2077 int i;
2078
2079 ecbdata.diff_words =
2080 xcalloc(1, sizeof(struct diff_words_data));
2081 ecbdata.diff_words->type = o->word_diff;
2082 ecbdata.diff_words->opt = o;
2083 if (!o->word_regex)
2084 o->word_regex = userdiff_word_regex(one);
2085 if (!o->word_regex)
2086 o->word_regex = userdiff_word_regex(two);
2087 if (!o->word_regex)
2088 o->word_regex = diff_word_regex_cfg;
2089 if (o->word_regex) {
2090 ecbdata.diff_words->word_regex = (regex_t *)
2091 xmalloc(sizeof(regex_t));
2092 if (regcomp(ecbdata.diff_words->word_regex,
2093 o->word_regex,
2094 REG_EXTENDED | REG_NEWLINE))
2095 die ("Invalid regular expression: %s",
2096 o->word_regex);
2097 }
2098 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2099 if (o->word_diff == diff_words_styles[i].type) {
2100 ecbdata.diff_words->style =
2101 &diff_words_styles[i];
2102 break;
2103 }
2104 }
2105 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2106 struct diff_words_style *st = ecbdata.diff_words->style;
2107 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2108 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2109 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2110 }
2111 }
2112 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2113 &xpp, &xecfg);
2114 if (o->word_diff)
2115 free_diff_words_data(&ecbdata);
2116 if (textconv_one)
2117 free(mf1.ptr);
2118 if (textconv_two)
2119 free(mf2.ptr);
2120 xdiff_clear_find_func(&xecfg);
2121 }
2122
2123 free_ab_and_return:
2124 strbuf_release(&header);
2125 diff_free_filespec_data(one);
2126 diff_free_filespec_data(two);
2127 free(a_one);
2128 free(b_two);
2129 return;
2130 }
2131
2132 static void builtin_diffstat(const char *name_a, const char *name_b,
2133 struct diff_filespec *one,
2134 struct diff_filespec *two,
2135 struct diffstat_t *diffstat,
2136 struct diff_options *o,
2137 int complete_rewrite)
2138 {
2139 mmfile_t mf1, mf2;
2140 struct diffstat_file *data;
2141
2142 data = diffstat_add(diffstat, name_a, name_b);
2143
2144 if (!one || !two) {
2145 data->is_unmerged = 1;
2146 return;
2147 }
2148
2149 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2150 data->is_binary = 1;
2151 data->added = diff_filespec_size(two);
2152 data->deleted = diff_filespec_size(one);
2153 }
2154
2155 else if (complete_rewrite) {
2156 diff_populate_filespec(one, 0);
2157 diff_populate_filespec(two, 0);
2158 data->deleted = count_lines(one->data, one->size);
2159 data->added = count_lines(two->data, two->size);
2160 }
2161
2162 else {
2163 /* Crazy xdl interfaces.. */
2164 xpparam_t xpp;
2165 xdemitconf_t xecfg;
2166
2167 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2168 die("unable to read files to diff");
2169
2170 memset(&xpp, 0, sizeof(xpp));
2171 memset(&xecfg, 0, sizeof(xecfg));
2172 xpp.flags = o->xdl_opts;
2173 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2174 &xpp, &xecfg);
2175 }
2176
2177 diff_free_filespec_data(one);
2178 diff_free_filespec_data(two);
2179 }
2180
2181 static void builtin_checkdiff(const char *name_a, const char *name_b,
2182 const char *attr_path,
2183 struct diff_filespec *one,
2184 struct diff_filespec *two,
2185 struct diff_options *o)
2186 {
2187 mmfile_t mf1, mf2;
2188 struct checkdiff_t data;
2189
2190 if (!two)
2191 return;
2192
2193 memset(&data, 0, sizeof(data));
2194 data.filename = name_b ? name_b : name_a;
2195 data.lineno = 0;
2196 data.o = o;
2197 data.ws_rule = whitespace_rule(attr_path);
2198 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2199
2200 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2201 die("unable to read files to diff");
2202
2203 /*
2204 * All the other codepaths check both sides, but not checking
2205 * the "old" side here is deliberate. We are checking the newly
2206 * introduced changes, and as long as the "new" side is text, we
2207 * can and should check what it introduces.
2208 */
2209 if (diff_filespec_is_binary(two))
2210 goto free_and_return;
2211 else {
2212 /* Crazy xdl interfaces.. */
2213 xpparam_t xpp;
2214 xdemitconf_t xecfg;
2215
2216 memset(&xpp, 0, sizeof(xpp));
2217 memset(&xecfg, 0, sizeof(xecfg));
2218 xecfg.ctxlen = 1; /* at least one context line */
2219 xpp.flags = 0;
2220 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2221 &xpp, &xecfg);
2222
2223 if (data.ws_rule & WS_BLANK_AT_EOF) {
2224 struct emit_callback ecbdata;
2225 int blank_at_eof;
2226
2227 ecbdata.ws_rule = data.ws_rule;
2228 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2229 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2230
2231 if (blank_at_eof) {
2232 static char *err;
2233 if (!err)
2234 err = whitespace_error_string(WS_BLANK_AT_EOF);
2235 fprintf(o->file, "%s:%d: %s.\n",
2236 data.filename, blank_at_eof, err);
2237 data.status = 1; /* report errors */
2238 }
2239 }
2240 }
2241 free_and_return:
2242 diff_free_filespec_data(one);
2243 diff_free_filespec_data(two);
2244 if (data.status)
2245 DIFF_OPT_SET(o, CHECK_FAILED);
2246 }
2247
2248 struct diff_filespec *alloc_filespec(const char *path)
2249 {
2250 int namelen = strlen(path);
2251 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2252
2253 memset(spec, 0, sizeof(*spec));
2254 spec->path = (char *)(spec + 1);
2255 memcpy(spec->path, path, namelen+1);
2256 spec->count = 1;
2257 spec->is_binary = -1;
2258 return spec;
2259 }
2260
2261 void free_filespec(struct diff_filespec *spec)
2262 {
2263 if (!--spec->count) {
2264 diff_free_filespec_data(spec);
2265 free(spec);
2266 }
2267 }
2268
2269 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2270 unsigned short mode)
2271 {
2272 if (mode) {
2273 spec->mode = canon_mode(mode);
2274 hashcpy(spec->sha1, sha1);
2275 spec->sha1_valid = !is_null_sha1(sha1);
2276 }
2277 }
2278
2279 /*
2280 * Given a name and sha1 pair, if the index tells us the file in
2281 * the work tree has that object contents, return true, so that
2282 * prepare_temp_file() does not have to inflate and extract.
2283 */
2284 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2285 {
2286 struct cache_entry *ce;
2287 struct stat st;
2288 int pos, len;
2289
2290 /*
2291 * We do not read the cache ourselves here, because the
2292 * benchmark with my previous version that always reads cache
2293 * shows that it makes things worse for diff-tree comparing
2294 * two linux-2.6 kernel trees in an already checked out work
2295 * tree. This is because most diff-tree comparisons deal with
2296 * only a small number of files, while reading the cache is
2297 * expensive for a large project, and its cost outweighs the
2298 * savings we get by not inflating the object to a temporary
2299 * file. Practically, this code only helps when we are used
2300 * by diff-cache --cached, which does read the cache before
2301 * calling us.
2302 */
2303 if (!active_cache)
2304 return 0;
2305
2306 /* We want to avoid the working directory if our caller
2307 * doesn't need the data in a normal file, this system
2308 * is rather slow with its stat/open/mmap/close syscalls,
2309 * and the object is contained in a pack file. The pack
2310 * is probably already open and will be faster to obtain
2311 * the data through than the working directory. Loose
2312 * objects however would tend to be slower as they need
2313 * to be individually opened and inflated.
2314 */
2315 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2316 return 0;
2317
2318 len = strlen(name);
2319 pos = cache_name_pos(name, len);
2320 if (pos < 0)
2321 return 0;
2322 ce = active_cache[pos];
2323
2324 /*
2325 * This is not the sha1 we are looking for, or
2326 * unreusable because it is not a regular file.
2327 */
2328 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2329 return 0;
2330
2331 /*
2332 * If ce is marked as "assume unchanged", there is no
2333 * guarantee that work tree matches what we are looking for.
2334 */
2335 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2336 return 0;
2337
2338 /*
2339 * If ce matches the file in the work tree, we can reuse it.
2340 */
2341 if (ce_uptodate(ce) ||
2342 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2343 return 1;
2344
2345 return 0;
2346 }
2347
2348 static int populate_from_stdin(struct diff_filespec *s)
2349 {
2350 struct strbuf buf = STRBUF_INIT;
2351 size_t size = 0;
2352
2353 if (strbuf_read(&buf, 0, 0) < 0)
2354 return error("error while reading from stdin %s",
2355 strerror(errno));
2356
2357 s->should_munmap = 0;
2358 s->data = strbuf_detach(&buf, &size);
2359 s->size = size;
2360 s->should_free = 1;
2361 return 0;
2362 }
2363
2364 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2365 {
2366 int len;
2367 char *data = xmalloc(100), *dirty = "";
2368
2369 /* Are we looking at the work tree? */
2370 if (s->dirty_submodule)
2371 dirty = "-dirty";
2372
2373 len = snprintf(data, 100,
2374 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2375 s->data = data;
2376 s->size = len;
2377 s->should_free = 1;
2378 if (size_only) {
2379 s->data = NULL;
2380 free(data);
2381 }
2382 return 0;
2383 }
2384
2385 /*
2386 * While doing rename detection and pickaxe operation, we may need to
2387 * grab the data for the blob (or file) for our own in-core comparison.
2388 * diff_filespec has data and size fields for this purpose.
2389 */
2390 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2391 {
2392 int err = 0;
2393 if (!DIFF_FILE_VALID(s))
2394 die("internal error: asking to populate invalid file.");
2395 if (S_ISDIR(s->mode))
2396 return -1;
2397
2398 if (s->data)
2399 return 0;
2400
2401 if (size_only && 0 < s->size)
2402 return 0;
2403
2404 if (S_ISGITLINK(s->mode))
2405 return diff_populate_gitlink(s, size_only);
2406
2407 if (!s->sha1_valid ||
2408 reuse_worktree_file(s->path, s->sha1, 0)) {
2409 struct strbuf buf = STRBUF_INIT;
2410 struct stat st;
2411 int fd;
2412
2413 if (!strcmp(s->path, "-"))
2414 return populate_from_stdin(s);
2415
2416 if (lstat(s->path, &st) < 0) {
2417 if (errno == ENOENT) {
2418 err_empty:
2419 err = -1;
2420 empty:
2421 s->data = (char *)"";
2422 s->size = 0;
2423 return err;
2424 }
2425 }
2426 s->size = xsize_t(st.st_size);
2427 if (!s->size)
2428 goto empty;
2429 if (S_ISLNK(st.st_mode)) {
2430 struct strbuf sb = STRBUF_INIT;
2431
2432 if (strbuf_readlink(&sb, s->path, s->size))
2433 goto err_empty;
2434 s->size = sb.len;
2435 s->data = strbuf_detach(&sb, NULL);
2436 s->should_free = 1;
2437 return 0;
2438 }
2439 if (size_only)
2440 return 0;
2441 fd = open(s->path, O_RDONLY);
2442 if (fd < 0)
2443 goto err_empty;
2444 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2445 close(fd);
2446 s->should_munmap = 1;
2447
2448 /*
2449 * Convert from working tree format to canonical git format
2450 */
2451 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2452 size_t size = 0;
2453 munmap(s->data, s->size);
2454 s->should_munmap = 0;
2455 s->data = strbuf_detach(&buf, &size);
2456 s->size = size;
2457 s->should_free = 1;
2458 }
2459 }
2460 else {
2461 enum object_type type;
2462 if (size_only) {
2463 type = sha1_object_info(s->sha1, &s->size);
2464 if (type < 0)
2465 die("unable to read %s", sha1_to_hex(s->sha1));
2466 } else {
2467 s->data = read_sha1_file(s->sha1, &type, &s->size);
2468 if (!s->data)
2469 die("unable to read %s", sha1_to_hex(s->sha1));
2470 s->should_free = 1;
2471 }
2472 }
2473 return 0;
2474 }
2475
2476 void diff_free_filespec_blob(struct diff_filespec *s)
2477 {
2478 if (s->should_free)
2479 free(s->data);
2480 else if (s->should_munmap)
2481 munmap(s->data, s->size);
2482
2483 if (s->should_free || s->should_munmap) {
2484 s->should_free = s->should_munmap = 0;
2485 s->data = NULL;
2486 }
2487 }
2488
2489 void diff_free_filespec_data(struct diff_filespec *s)
2490 {
2491 diff_free_filespec_blob(s);
2492 free(s->cnt_data);
2493 s->cnt_data = NULL;
2494 }
2495
2496 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2497 void *blob,
2498 unsigned long size,
2499 const unsigned char *sha1,
2500 int mode)
2501 {
2502 int fd;
2503 struct strbuf buf = STRBUF_INIT;
2504 struct strbuf template = STRBUF_INIT;
2505 char *path_dup = xstrdup(path);
2506 const char *base = basename(path_dup);
2507
2508 /* Generate "XXXXXX_basename.ext" */
2509 strbuf_addstr(&template, "XXXXXX_");
2510 strbuf_addstr(&template, base);
2511
2512 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2513 strlen(base) + 1);
2514 if (fd < 0)
2515 die_errno("unable to create temp-file");
2516 if (convert_to_working_tree(path,
2517 (const char *)blob, (size_t)size, &buf)) {
2518 blob = buf.buf;
2519 size = buf.len;
2520 }
2521 if (write_in_full(fd, blob, size) != size)
2522 die_errno("unable to write temp-file");
2523 close(fd);
2524 temp->name = temp->tmp_path;
2525 strcpy(temp->hex, sha1_to_hex(sha1));
2526 temp->hex[40] = 0;
2527 sprintf(temp->mode, "%06o", mode);
2528 strbuf_release(&buf);
2529 strbuf_release(&template);
2530 free(path_dup);
2531 }
2532
2533 static struct diff_tempfile *prepare_temp_file(const char *name,
2534 struct diff_filespec *one)
2535 {
2536 struct diff_tempfile *temp = claim_diff_tempfile();
2537
2538 if (!DIFF_FILE_VALID(one)) {
2539 not_a_valid_file:
2540 /* A '-' entry produces this for file-2, and
2541 * a '+' entry produces this for file-1.
2542 */
2543 temp->name = "/dev/null";
2544 strcpy(temp->hex, ".");
2545 strcpy(temp->mode, ".");
2546 return temp;
2547 }
2548
2549 if (!remove_tempfile_installed) {
2550 atexit(remove_tempfile);
2551 sigchain_push_common(remove_tempfile_on_signal);
2552 remove_tempfile_installed = 1;
2553 }
2554
2555 if (!one->sha1_valid ||
2556 reuse_worktree_file(name, one->sha1, 1)) {
2557 struct stat st;
2558 if (lstat(name, &st) < 0) {
2559 if (errno == ENOENT)
2560 goto not_a_valid_file;
2561 die_errno("stat(%s)", name);
2562 }
2563 if (S_ISLNK(st.st_mode)) {
2564 struct strbuf sb = STRBUF_INIT;
2565 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2566 die_errno("readlink(%s)", name);
2567 prep_temp_blob(name, temp, sb.buf, sb.len,
2568 (one->sha1_valid ?
2569 one->sha1 : null_sha1),
2570 (one->sha1_valid ?
2571 one->mode : S_IFLNK));
2572 strbuf_release(&sb);
2573 }
2574 else {
2575 /* we can borrow from the file in the work tree */
2576 temp->name = name;
2577 if (!one->sha1_valid)
2578 strcpy(temp->hex, sha1_to_hex(null_sha1));
2579 else
2580 strcpy(temp->hex, sha1_to_hex(one->sha1));
2581 /* Even though we may sometimes borrow the
2582 * contents from the work tree, we always want
2583 * one->mode. mode is trustworthy even when
2584 * !(one->sha1_valid), as long as
2585 * DIFF_FILE_VALID(one).
2586 */
2587 sprintf(temp->mode, "%06o", one->mode);
2588 }
2589 return temp;
2590 }
2591 else {
2592 if (diff_populate_filespec(one, 0))
2593 die("cannot read data blob for %s", one->path);
2594 prep_temp_blob(name, temp, one->data, one->size,
2595 one->sha1, one->mode);
2596 }
2597 return temp;
2598 }
2599
2600 /* An external diff command takes:
2601 *
2602 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2603 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2604 *
2605 */
2606 static void run_external_diff(const char *pgm,
2607 const char *name,
2608 const char *other,
2609 struct diff_filespec *one,
2610 struct diff_filespec *two,
2611 const char *xfrm_msg,
2612 int complete_rewrite)
2613 {
2614 const char *spawn_arg[10];
2615 int retval;
2616 const char **arg = &spawn_arg[0];
2617
2618 if (one && two) {
2619 struct diff_tempfile *temp_one, *temp_two;
2620 const char *othername = (other ? other : name);
2621 temp_one = prepare_temp_file(name, one);
2622 temp_two = prepare_temp_file(othername, two);
2623 *arg++ = pgm;
2624 *arg++ = name;
2625 *arg++ = temp_one->name;
2626 *arg++ = temp_one->hex;
2627 *arg++ = temp_one->mode;
2628 *arg++ = temp_two->name;
2629 *arg++ = temp_two->hex;
2630 *arg++ = temp_two->mode;
2631 if (other) {
2632 *arg++ = other;
2633 *arg++ = xfrm_msg;
2634 }
2635 } else {
2636 *arg++ = pgm;
2637 *arg++ = name;
2638 }
2639 *arg = NULL;
2640 fflush(NULL);
2641 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2642 remove_tempfile();
2643 if (retval) {
2644 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2645 exit(1);
2646 }
2647 }
2648
2649 static int similarity_index(struct diff_filepair *p)
2650 {
2651 return p->score * 100 / MAX_SCORE;
2652 }
2653
2654 static void fill_metainfo(struct strbuf *msg,
2655 const char *name,
2656 const char *other,
2657 struct diff_filespec *one,
2658 struct diff_filespec *two,
2659 struct diff_options *o,
2660 struct diff_filepair *p,
2661 int *must_show_header,
2662 int use_color)
2663 {
2664 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2665 const char *reset = diff_get_color(use_color, DIFF_RESET);
2666 struct strbuf *msgbuf;
2667 char *line_prefix = "";
2668
2669 *must_show_header = 1;
2670 if (o->output_prefix) {
2671 msgbuf = o->output_prefix(o, o->output_prefix_data);
2672 line_prefix = msgbuf->buf;
2673 }
2674 strbuf_init(msg, PATH_MAX * 2 + 300);
2675 switch (p->status) {
2676 case DIFF_STATUS_COPIED:
2677 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2678 line_prefix, set, similarity_index(p));
2679 strbuf_addf(msg, "%s\n%s%scopy from ",
2680 reset, line_prefix, set);
2681 quote_c_style(name, msg, NULL, 0);
2682 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2683 quote_c_style(other, msg, NULL, 0);
2684 strbuf_addf(msg, "%s\n", reset);
2685 break;
2686 case DIFF_STATUS_RENAMED:
2687 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2688 line_prefix, set, similarity_index(p));
2689 strbuf_addf(msg, "%s\n%s%srename from ",
2690 reset, line_prefix, set);
2691 quote_c_style(name, msg, NULL, 0);
2692 strbuf_addf(msg, "%s\n%s%srename to ",
2693 reset, line_prefix, set);
2694 quote_c_style(other, msg, NULL, 0);
2695 strbuf_addf(msg, "%s\n", reset);
2696 break;
2697 case DIFF_STATUS_MODIFIED:
2698 if (p->score) {
2699 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2700 line_prefix,
2701 set, similarity_index(p), reset);
2702 break;
2703 }
2704 /* fallthru */
2705 default:
2706 *must_show_header = 0;
2707 }
2708 if (one && two && hashcmp(one->sha1, two->sha1)) {
2709 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2710
2711 if (DIFF_OPT_TST(o, BINARY)) {
2712 mmfile_t mf;
2713 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2714 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2715 abbrev = 40;
2716 }
2717 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2718 find_unique_abbrev(one->sha1, abbrev));
2719 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2720 if (one->mode == two->mode)
2721 strbuf_addf(msg, " %06o", one->mode);
2722 strbuf_addf(msg, "%s\n", reset);
2723 }
2724 }
2725
2726 static void run_diff_cmd(const char *pgm,
2727 const char *name,
2728 const char *other,
2729 const char *attr_path,
2730 struct diff_filespec *one,
2731 struct diff_filespec *two,
2732 struct strbuf *msg,
2733 struct diff_options *o,
2734 struct diff_filepair *p)
2735 {
2736 const char *xfrm_msg = NULL;
2737 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2738 int must_show_header = 0;
2739
2740 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2741 pgm = NULL;
2742 else {
2743 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2744 if (drv && drv->external)
2745 pgm = drv->external;
2746 }
2747
2748 if (msg) {
2749 /*
2750 * don't use colors when the header is intended for an
2751 * external diff driver
2752 */
2753 fill_metainfo(msg, name, other, one, two, o, p,
2754 &must_show_header,
2755 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2756 xfrm_msg = msg->len ? msg->buf : NULL;
2757 }
2758
2759 if (pgm) {
2760 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2761 complete_rewrite);
2762 return;
2763 }
2764 if (one && two)
2765 builtin_diff(name, other ? other : name,
2766 one, two, xfrm_msg, must_show_header,
2767 o, complete_rewrite);
2768 else
2769 fprintf(o->file, "* Unmerged path %s\n", name);
2770 }
2771
2772 static void diff_fill_sha1_info(struct diff_filespec *one)
2773 {
2774 if (DIFF_FILE_VALID(one)) {
2775 if (!one->sha1_valid) {
2776 struct stat st;
2777 if (!strcmp(one->path, "-")) {
2778 hashcpy(one->sha1, null_sha1);
2779 return;
2780 }
2781 if (lstat(one->path, &st) < 0)
2782 die_errno("stat '%s'", one->path);
2783 if (index_path(one->sha1, one->path, &st, 0))
2784 die("cannot hash %s", one->path);
2785 }
2786 }
2787 else
2788 hashclr(one->sha1);
2789 }
2790
2791 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2792 {
2793 /* Strip the prefix but do not molest /dev/null and absolute paths */
2794 if (*namep && **namep != '/') {
2795 *namep += prefix_length;
2796 if (**namep == '/')
2797 ++*namep;
2798 }
2799 if (*otherp && **otherp != '/') {
2800 *otherp += prefix_length;
2801 if (**otherp == '/')
2802 ++*otherp;
2803 }
2804 }
2805
2806 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2807 {
2808 const char *pgm = external_diff();
2809 struct strbuf msg;
2810 struct diff_filespec *one = p->one;
2811 struct diff_filespec *two = p->two;
2812 const char *name;
2813 const char *other;
2814 const char *attr_path;
2815
2816 name = p->one->path;
2817 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2818 attr_path = name;
2819 if (o->prefix_length)
2820 strip_prefix(o->prefix_length, &name, &other);
2821
2822 if (DIFF_PAIR_UNMERGED(p)) {
2823 run_diff_cmd(pgm, name, NULL, attr_path,
2824 NULL, NULL, NULL, o, p);
2825 return;
2826 }
2827
2828 diff_fill_sha1_info(one);
2829 diff_fill_sha1_info(two);
2830
2831 if (!pgm &&
2832 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2833 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2834 /*
2835 * a filepair that changes between file and symlink
2836 * needs to be split into deletion and creation.
2837 */
2838 struct diff_filespec *null = alloc_filespec(two->path);
2839 run_diff_cmd(NULL, name, other, attr_path,
2840 one, null, &msg, o, p);
2841 free(null);
2842 strbuf_release(&msg);
2843
2844 null = alloc_filespec(one->path);
2845 run_diff_cmd(NULL, name, other, attr_path,
2846 null, two, &msg, o, p);
2847 free(null);
2848 }
2849 else
2850 run_diff_cmd(pgm, name, other, attr_path,
2851 one, two, &msg, o, p);
2852
2853 strbuf_release(&msg);
2854 }
2855
2856 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2857 struct diffstat_t *diffstat)
2858 {
2859 const char *name;
2860 const char *other;
2861 int complete_rewrite = 0;
2862
2863 if (DIFF_PAIR_UNMERGED(p)) {
2864 /* unmerged */
2865 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2866 return;
2867 }
2868
2869 name = p->one->path;
2870 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2871
2872 if (o->prefix_length)
2873 strip_prefix(o->prefix_length, &name, &other);
2874
2875 diff_fill_sha1_info(p->one);
2876 diff_fill_sha1_info(p->two);
2877
2878 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2879 complete_rewrite = 1;
2880 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2881 }
2882
2883 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2884 {
2885 const char *name;
2886 const char *other;
2887 const char *attr_path;
2888
2889 if (DIFF_PAIR_UNMERGED(p)) {
2890 /* unmerged */
2891 return;
2892 }
2893
2894 name = p->one->path;
2895 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2896 attr_path = other ? other : name;
2897
2898 if (o->prefix_length)
2899 strip_prefix(o->prefix_length, &name, &other);
2900
2901 diff_fill_sha1_info(p->one);
2902 diff_fill_sha1_info(p->two);
2903
2904 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2905 }
2906
2907 void diff_setup(struct diff_options *options)
2908 {
2909 memcpy(options, &default_diff_options, sizeof(*options));
2910
2911 options->file = stdout;
2912
2913 options->line_termination = '\n';
2914 options->break_opt = -1;
2915 options->rename_limit = -1;
2916 options->dirstat_percent = 3;
2917 options->context = 3;
2918
2919 options->change = diff_change;
2920 options->add_remove = diff_addremove;
2921 if (diff_use_color_default > 0)
2922 DIFF_OPT_SET(options, COLOR_DIFF);
2923 options->detect_rename = diff_detect_rename_default;
2924
2925 if (diff_no_prefix) {
2926 options->a_prefix = options->b_prefix = "";
2927 } else if (!diff_mnemonic_prefix) {
2928 options->a_prefix = "a/";
2929 options->b_prefix = "b/";
2930 }
2931 }
2932
2933 int diff_setup_done(struct diff_options *options)
2934 {
2935 int count = 0;
2936
2937 if (options->output_format & DIFF_FORMAT_NAME)
2938 count++;
2939 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2940 count++;
2941 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2942 count++;
2943 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2944 count++;
2945 if (count > 1)
2946 die("--name-only, --name-status, --check and -s are mutually exclusive");
2947
2948 /*
2949 * Most of the time we can say "there are changes"
2950 * only by checking if there are changed paths, but
2951 * --ignore-whitespace* options force us to look
2952 * inside contents.
2953 */
2954
2955 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2956 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2957 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2958 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2959 else
2960 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2961
2962 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2963 options->detect_rename = DIFF_DETECT_COPY;
2964
2965 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2966 options->prefix = NULL;
2967 if (options->prefix)
2968 options->prefix_length = strlen(options->prefix);
2969 else
2970 options->prefix_length = 0;
2971
2972 if (options->output_format & (DIFF_FORMAT_NAME |
2973 DIFF_FORMAT_NAME_STATUS |
2974 DIFF_FORMAT_CHECKDIFF |
2975 DIFF_FORMAT_NO_OUTPUT))
2976 options->output_format &= ~(DIFF_FORMAT_RAW |
2977 DIFF_FORMAT_NUMSTAT |
2978 DIFF_FORMAT_DIFFSTAT |
2979 DIFF_FORMAT_SHORTSTAT |
2980 DIFF_FORMAT_DIRSTAT |
2981 DIFF_FORMAT_SUMMARY |
2982 DIFF_FORMAT_PATCH);
2983
2984 /*
2985 * These cases always need recursive; we do not drop caller-supplied
2986 * recursive bits for other formats here.
2987 */
2988 if (options->output_format & (DIFF_FORMAT_PATCH |
2989 DIFF_FORMAT_NUMSTAT |
2990 DIFF_FORMAT_DIFFSTAT |
2991 DIFF_FORMAT_SHORTSTAT |
2992 DIFF_FORMAT_DIRSTAT |
2993 DIFF_FORMAT_SUMMARY |
2994 DIFF_FORMAT_CHECKDIFF))
2995 DIFF_OPT_SET(options, RECURSIVE);
2996 /*
2997 * Also pickaxe would not work very well if you do not say recursive
2998 */
2999 if (options->pickaxe)
3000 DIFF_OPT_SET(options, RECURSIVE);
3001 /*
3002 * When patches are generated, submodules diffed against the work tree
3003 * must be checked for dirtiness too so it can be shown in the output
3004 */
3005 if (options->output_format & DIFF_FORMAT_PATCH)
3006 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3007
3008 if (options->detect_rename && options->rename_limit < 0)
3009 options->rename_limit = diff_rename_limit_default;
3010 if (options->setup & DIFF_SETUP_USE_CACHE) {
3011 if (!active_cache)
3012 /* read-cache does not die even when it fails
3013 * so it is safe for us to do this here. Also
3014 * it does not smudge active_cache or active_nr
3015 * when it fails, so we do not have to worry about
3016 * cleaning it up ourselves either.
3017 */
3018 read_cache();
3019 }
3020 if (options->abbrev <= 0 || 40 < options->abbrev)
3021 options->abbrev = 40; /* full */
3022
3023 /*
3024 * It does not make sense to show the first hit we happened
3025 * to have found. It does not make sense not to return with
3026 * exit code in such a case either.
3027 */
3028 if (DIFF_OPT_TST(options, QUICK)) {
3029 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3030 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3031 }
3032
3033 return 0;
3034 }
3035
3036 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3037 {
3038 char c, *eq;
3039 int len;
3040
3041 if (*arg != '-')
3042 return 0;
3043 c = *++arg;
3044 if (!c)
3045 return 0;
3046 if (c == arg_short) {
3047 c = *++arg;
3048 if (!c)
3049 return 1;
3050 if (val && isdigit(c)) {
3051 char *end;
3052 int n = strtoul(arg, &end, 10);
3053 if (*end)
3054 return 0;
3055 *val = n;
3056 return 1;
3057 }
3058 return 0;
3059 }
3060 if (c != '-')
3061 return 0;
3062 arg++;
3063 eq = strchr(arg, '=');
3064 if (eq)
3065 len = eq - arg;
3066 else
3067 len = strlen(arg);
3068 if (!len || strncmp(arg, arg_long, len))
3069 return 0;
3070 if (eq) {
3071 int n;
3072 char *end;
3073 if (!isdigit(*++eq))
3074 return 0;
3075 n = strtoul(eq, &end, 10);
3076 if (*end)
3077 return 0;
3078 *val = n;
3079 }
3080 return 1;
3081 }
3082
3083 static int diff_scoreopt_parse(const char *opt);
3084
3085 static inline int short_opt(char opt, const char **argv,
3086 const char **optarg)
3087 {
3088 const char *arg = argv[0];
3089 if (arg[0] != '-' || arg[1] != opt)
3090 return 0;
3091 if (arg[2] != '\0') {
3092 *optarg = arg + 2;
3093 return 1;
3094 }
3095 if (!argv[1])
3096 die("Option '%c' requires a value", opt);
3097 *optarg = argv[1];
3098 return 2;
3099 }
3100
3101 int parse_long_opt(const char *opt, const char **argv,
3102 const char **optarg)
3103 {
3104 const char *arg = argv[0];
3105 if (arg[0] != '-' || arg[1] != '-')
3106 return 0;
3107 arg += strlen("--");
3108 if (prefixcmp(arg, opt))
3109 return 0;
3110 arg += strlen(opt);
3111 if (*arg == '=') { /* sticked form: --option=value */
3112 *optarg = arg + 1;
3113 return 1;
3114 }
3115 if (*arg != '\0')
3116 return 0;
3117 /* separate form: --option value */
3118 if (!argv[1])
3119 die("Option '--%s' requires a value", opt);
3120 *optarg = argv[1];
3121 return 2;
3122 }
3123
3124 static int stat_opt(struct diff_options *options, const char **av)
3125 {
3126 const char *arg = av[0];
3127 char *end;
3128 int width = options->stat_width;
3129 int name_width = options->stat_name_width;
3130 int count = options->stat_count;
3131 int argcount = 1;
3132
3133 arg += strlen("--stat");
3134 end = (char *)arg;
3135
3136 switch (*arg) {
3137 case '-':
3138 if (!prefixcmp(arg, "-width")) {
3139 arg += strlen("-width");
3140 if (*arg == '=')
3141 width = strtoul(arg + 1, &end, 10);
3142 else if (!*arg && !av[1])
3143 die("Option '--stat-width' requires a value");
3144 else if (!*arg) {
3145 width = strtoul(av[1], &end, 10);
3146 argcount = 2;
3147 }
3148 } else if (!prefixcmp(arg, "-name-width")) {
3149 arg += strlen("-name-width");
3150 if (*arg == '=')
3151 name_width = strtoul(arg + 1, &end, 10);
3152 else if (!*arg && !av[1])
3153 die("Option '--stat-name-width' requires a value");
3154 else if (!*arg) {
3155 name_width = strtoul(av[1], &end, 10);
3156 argcount = 2;
3157 }
3158 } else if (!prefixcmp(arg, "-count")) {
3159 arg += strlen("-count");
3160 if (*arg == '=')
3161 count = strtoul(arg + 1, &end, 10);
3162 else if (!*arg && !av[1])
3163 die("Option '--stat-count' requires a value");
3164 else if (!*arg) {
3165 count = strtoul(av[1], &end, 10);
3166 argcount = 2;
3167 }
3168 }
3169 break;
3170 case '=':
3171 width = strtoul(arg+1, &end, 10);
3172 if (*end == ',')
3173 name_width = strtoul(end+1, &end, 10);
3174 if (*end == ',')
3175 count = strtoul(end+1, &end, 10);
3176 }
3177
3178 /* Important! This checks all the error cases! */
3179 if (*end)
3180 return 0;
3181 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3182 options->stat_name_width = name_width;
3183 options->stat_width = width;
3184 options->stat_count = count;
3185 return argcount;
3186 }
3187
3188 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3189 {
3190 const char *arg = av[0];
3191 const char *optarg;
3192 int argcount;
3193
3194 /* Output format options */
3195 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3196 options->output_format |= DIFF_FORMAT_PATCH;
3197 else if (opt_arg(arg, 'U', "unified", &options->context))
3198 options->output_format |= DIFF_FORMAT_PATCH;
3199 else if (!strcmp(arg, "--raw"))
3200 options->output_format |= DIFF_FORMAT_RAW;
3201 else if (!strcmp(arg, "--patch-with-raw"))
3202 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3203 else if (!strcmp(arg, "--numstat"))
3204 options->output_format |= DIFF_FORMAT_NUMSTAT;
3205 else if (!strcmp(arg, "--shortstat"))
3206 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3207 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3208 options->output_format |= DIFF_FORMAT_DIRSTAT;
3209 else if (!strcmp(arg, "--cumulative")) {
3210 options->output_format |= DIFF_FORMAT_DIRSTAT;
3211 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3212 } else if (opt_arg(arg, 0, "dirstat-by-file",
3213 &options->dirstat_percent)) {
3214 options->output_format |= DIFF_FORMAT_DIRSTAT;
3215 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3216 }
3217 else if (!strcmp(arg, "--check"))
3218 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3219 else if (!strcmp(arg, "--summary"))
3220 options->output_format |= DIFF_FORMAT_SUMMARY;
3221 else if (!strcmp(arg, "--patch-with-stat"))
3222 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3223 else if (!strcmp(arg, "--name-only"))
3224 options->output_format |= DIFF_FORMAT_NAME;
3225 else if (!strcmp(arg, "--name-status"))
3226 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3227 else if (!strcmp(arg, "-s"))
3228 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3229 else if (!prefixcmp(arg, "--stat"))
3230 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3231 return stat_opt(options, av);
3232
3233 /* renames options */
3234 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3235 !strcmp(arg, "--break-rewrites")) {
3236 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3237 return error("invalid argument to -B: %s", arg+2);
3238 }
3239 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3240 !strcmp(arg, "--find-renames")) {
3241 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3242 return error("invalid argument to -M: %s", arg+2);
3243 options->detect_rename = DIFF_DETECT_RENAME;
3244 }
3245 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3246 options->irreversible_delete = 1;
3247 }
3248 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3249 !strcmp(arg, "--find-copies")) {
3250 if (options->detect_rename == DIFF_DETECT_COPY)
3251 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3252 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3253 return error("invalid argument to -C: %s", arg+2);
3254 options->detect_rename = DIFF_DETECT_COPY;
3255 }
3256 else if (!strcmp(arg, "--no-renames"))
3257 options->detect_rename = 0;
3258 else if (!strcmp(arg, "--relative"))
3259 DIFF_OPT_SET(options, RELATIVE_NAME);
3260 else if (!prefixcmp(arg, "--relative=")) {
3261 DIFF_OPT_SET(options, RELATIVE_NAME);
3262 options->prefix = arg + 11;
3263 }
3264
3265 /* xdiff options */
3266 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3267 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3268 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3269 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3270 else if (!strcmp(arg, "--ignore-space-at-eol"))
3271 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3272 else if (!strcmp(arg, "--patience"))
3273 DIFF_XDL_SET(options, PATIENCE_DIFF);
3274
3275 /* flags options */
3276 else if (!strcmp(arg, "--binary")) {
3277 options->output_format |= DIFF_FORMAT_PATCH;
3278 DIFF_OPT_SET(options, BINARY);
3279 }
3280 else if (!strcmp(arg, "--full-index"))
3281 DIFF_OPT_SET(options, FULL_INDEX);
3282 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3283 DIFF_OPT_SET(options, TEXT);
3284 else if (!strcmp(arg, "-R"))
3285 DIFF_OPT_SET(options, REVERSE_DIFF);
3286 else if (!strcmp(arg, "--find-copies-harder"))
3287 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3288 else if (!strcmp(arg, "--follow"))
3289 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3290 else if (!strcmp(arg, "--color"))
3291 DIFF_OPT_SET(options, COLOR_DIFF);
3292 else if (!prefixcmp(arg, "--color=")) {
3293 int value = git_config_colorbool(NULL, arg+8, -1);
3294 if (value == 0)
3295 DIFF_OPT_CLR(options, COLOR_DIFF);
3296 else if (value > 0)
3297 DIFF_OPT_SET(options, COLOR_DIFF);
3298 else
3299 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3300 }
3301 else if (!strcmp(arg, "--no-color"))
3302 DIFF_OPT_CLR(options, COLOR_DIFF);
3303 else if (!strcmp(arg, "--color-words")) {
3304 DIFF_OPT_SET(options, COLOR_DIFF);
3305 options->word_diff = DIFF_WORDS_COLOR;
3306 }
3307 else if (!prefixcmp(arg, "--color-words=")) {
3308 DIFF_OPT_SET(options, COLOR_DIFF);
3309 options->word_diff = DIFF_WORDS_COLOR;
3310 options->word_regex = arg + 14;
3311 }
3312 else if (!strcmp(arg, "--word-diff")) {
3313 if (options->word_diff == DIFF_WORDS_NONE)
3314 options->word_diff = DIFF_WORDS_PLAIN;
3315 }
3316 else if (!prefixcmp(arg, "--word-diff=")) {
3317 const char *type = arg + 12;
3318 if (!strcmp(type, "plain"))
3319 options->word_diff = DIFF_WORDS_PLAIN;
3320 else if (!strcmp(type, "color")) {
3321 DIFF_OPT_SET(options, COLOR_DIFF);
3322 options->word_diff = DIFF_WORDS_COLOR;
3323 }
3324 else if (!strcmp(type, "porcelain"))
3325 options->word_diff = DIFF_WORDS_PORCELAIN;
3326 else if (!strcmp(type, "none"))
3327 options->word_diff = DIFF_WORDS_NONE;
3328 else
3329 die("bad --word-diff argument: %s", type);
3330 }
3331 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3332 if (options->word_diff == DIFF_WORDS_NONE)
3333 options->word_diff = DIFF_WORDS_PLAIN;
3334 options->word_regex = optarg;
3335 return argcount;
3336 }
3337 else if (!strcmp(arg, "--exit-code"))
3338 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3339 else if (!strcmp(arg, "--quiet"))
3340 DIFF_OPT_SET(options, QUICK);
3341 else if (!strcmp(arg, "--ext-diff"))
3342 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3343 else if (!strcmp(arg, "--no-ext-diff"))
3344 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3345 else if (!strcmp(arg, "--textconv"))
3346 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3347 else if (!strcmp(arg, "--no-textconv"))
3348 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3349 else if (!strcmp(arg, "--ignore-submodules")) {
3350 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3351 handle_ignore_submodules_arg(options, "all");
3352 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3353 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3354 handle_ignore_submodules_arg(options, arg + 20);
3355 } else if (!strcmp(arg, "--submodule"))
3356 DIFF_OPT_SET(options, SUBMODULE_LOG);
3357 else if (!prefixcmp(arg, "--submodule=")) {
3358 if (!strcmp(arg + 12, "log"))
3359 DIFF_OPT_SET(options, SUBMODULE_LOG);
3360 }
3361
3362 /* misc options */
3363 else if (!strcmp(arg, "-z"))
3364 options->line_termination = 0;
3365 else if ((argcount = short_opt('l', av, &optarg))) {
3366 options->rename_limit = strtoul(optarg, NULL, 10);
3367 return argcount;
3368 }
3369 else if ((argcount = short_opt('S', av, &optarg))) {
3370 options->pickaxe = optarg;
3371 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3372 return argcount;
3373 } else if ((argcount = short_opt('G', av, &optarg))) {
3374 options->pickaxe = optarg;
3375 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3376 return argcount;
3377 }
3378 else if (!strcmp(arg, "--pickaxe-all"))
3379 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3380 else if (!strcmp(arg, "--pickaxe-regex"))
3381 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3382 else if ((argcount = short_opt('O', av, &optarg))) {
3383 options->orderfile = optarg;
3384 return argcount;
3385 }
3386 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3387 options->filter = optarg;
3388 return argcount;
3389 }
3390 else if (!strcmp(arg, "--abbrev"))
3391 options->abbrev = DEFAULT_ABBREV;
3392 else if (!prefixcmp(arg, "--abbrev=")) {
3393 options->abbrev = strtoul(arg + 9, NULL, 10);
3394 if (options->abbrev < MINIMUM_ABBREV)
3395 options->abbrev = MINIMUM_ABBREV;
3396 else if (40 < options->abbrev)
3397 options->abbrev = 40;
3398 }
3399 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3400 options->a_prefix = optarg;
3401 return argcount;
3402 }
3403 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3404 options->b_prefix = optarg;
3405 return argcount;
3406 }
3407 else if (!strcmp(arg, "--no-prefix"))
3408 options->a_prefix = options->b_prefix = "";
3409 else if (opt_arg(arg, '\0', "inter-hunk-context",
3410 &options->interhunkcontext))
3411 ;
3412 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3413 options->file = fopen(optarg, "w");
3414 if (!options->file)
3415 die_errno("Could not open '%s'", optarg);
3416 options->close_file = 1;
3417 return argcount;
3418 } else
3419 return 0;
3420 return 1;
3421 }
3422
3423 int parse_rename_score(const char **cp_p)
3424 {
3425 unsigned long num, scale;
3426 int ch, dot;
3427 const char *cp = *cp_p;
3428
3429 num = 0;
3430 scale = 1;
3431 dot = 0;
3432 for (;;) {
3433 ch = *cp;
3434 if ( !dot && ch == '.' ) {
3435 scale = 1;
3436 dot = 1;
3437 } else if ( ch == '%' ) {
3438 scale = dot ? scale*100 : 100;
3439 cp++; /* % is always at the end */
3440 break;
3441 } else if ( ch >= '0' && ch <= '9' ) {
3442 if ( scale < 100000 ) {
3443 scale *= 10;
3444 num = (num*10) + (ch-'0');
3445 }
3446 } else {
3447 break;
3448 }
3449 cp++;
3450 }
3451 *cp_p = cp;
3452
3453 /* user says num divided by scale and we say internally that
3454 * is MAX_SCORE * num / scale.
3455 */
3456 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3457 }
3458
3459 static int diff_scoreopt_parse(const char *opt)
3460 {
3461 int opt1, opt2, cmd;
3462
3463 if (*opt++ != '-')
3464 return -1;
3465 cmd = *opt++;
3466 if (cmd == '-') {
3467 /* convert the long-form arguments into short-form versions */
3468 if (!prefixcmp(opt, "break-rewrites")) {
3469 opt += strlen("break-rewrites");
3470 if (*opt == 0 || *opt++ == '=')
3471 cmd = 'B';
3472 } else if (!prefixcmp(opt, "find-copies")) {
3473 opt += strlen("find-copies");
3474 if (*opt == 0 || *opt++ == '=')
3475 cmd = 'C';
3476 } else if (!prefixcmp(opt, "find-renames")) {
3477 opt += strlen("find-renames");
3478 if (*opt == 0 || *opt++ == '=')
3479 cmd = 'M';
3480 }
3481 }
3482 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3483 return -1; /* that is not a -M, -C nor -B option */
3484
3485 opt1 = parse_rename_score(&opt);
3486 if (cmd != 'B')
3487 opt2 = 0;
3488 else {
3489 if (*opt == 0)
3490 opt2 = 0;
3491 else if (*opt != '/')
3492 return -1; /* we expect -B80/99 or -B80 */
3493 else {
3494 opt++;
3495 opt2 = parse_rename_score(&opt);
3496 }
3497 }
3498 if (*opt != 0)
3499 return -1;
3500 return opt1 | (opt2 << 16);
3501 }
3502
3503 struct diff_queue_struct diff_queued_diff;
3504
3505 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3506 {
3507 if (queue->alloc <= queue->nr) {
3508 queue->alloc = alloc_nr(queue->alloc);
3509 queue->queue = xrealloc(queue->queue,
3510 sizeof(dp) * queue->alloc);
3511 }
3512 queue->queue[queue->nr++] = dp;
3513 }
3514
3515 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3516 struct diff_filespec *one,
3517 struct diff_filespec *two)
3518 {
3519 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3520 dp->one = one;
3521 dp->two = two;
3522 if (queue)
3523 diff_q(queue, dp);
3524 return dp;
3525 }
3526
3527 void diff_free_filepair(struct diff_filepair *p)
3528 {
3529 free_filespec(p->one);
3530 free_filespec(p->two);
3531 free(p);
3532 }
3533
3534 /* This is different from find_unique_abbrev() in that
3535 * it stuffs the result with dots for alignment.
3536 */
3537 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3538 {
3539 int abblen;
3540 const char *abbrev;
3541 if (len == 40)
3542 return sha1_to_hex(sha1);
3543
3544 abbrev = find_unique_abbrev(sha1, len);
3545 abblen = strlen(abbrev);
3546 if (abblen < 37) {
3547 static char hex[41];
3548 if (len < abblen && abblen <= len + 2)
3549 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3550 else
3551 sprintf(hex, "%s...", abbrev);
3552 return hex;
3553 }
3554 return sha1_to_hex(sha1);
3555 }
3556
3557 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3558 {
3559 int line_termination = opt->line_termination;
3560 int inter_name_termination = line_termination ? '\t' : '\0';
3561 if (opt->output_prefix) {
3562 struct strbuf *msg = NULL;
3563 msg = opt->output_prefix(opt, opt->output_prefix_data);
3564 fprintf(opt->file, "%s", msg->buf);
3565 }
3566
3567 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3568 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3569 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3570 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3571 }
3572 if (p->score) {
3573 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3574 inter_name_termination);
3575 } else {
3576 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3577 }
3578
3579 if (p->status == DIFF_STATUS_COPIED ||
3580 p->status == DIFF_STATUS_RENAMED) {
3581 const char *name_a, *name_b;
3582 name_a = p->one->path;
3583 name_b = p->two->path;
3584 strip_prefix(opt->prefix_length, &name_a, &name_b);
3585 write_name_quoted(name_a, opt->file, inter_name_termination);
3586 write_name_quoted(name_b, opt->file, line_termination);
3587 } else {
3588 const char *name_a, *name_b;
3589 name_a = p->one->mode ? p->one->path : p->two->path;
3590 name_b = NULL;
3591 strip_prefix(opt->prefix_length, &name_a, &name_b);
3592 write_name_quoted(name_a, opt->file, line_termination);
3593 }
3594 }
3595
3596 int diff_unmodified_pair(struct diff_filepair *p)
3597 {
3598 /* This function is written stricter than necessary to support
3599 * the currently implemented transformers, but the idea is to
3600 * let transformers to produce diff_filepairs any way they want,
3601 * and filter and clean them up here before producing the output.
3602 */
3603 struct diff_filespec *one = p->one, *two = p->two;
3604
3605 if (DIFF_PAIR_UNMERGED(p))
3606 return 0; /* unmerged is interesting */
3607
3608 /* deletion, addition, mode or type change
3609 * and rename are all interesting.
3610 */
3611 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3612 DIFF_PAIR_MODE_CHANGED(p) ||
3613 strcmp(one->path, two->path))
3614 return 0;
3615
3616 /* both are valid and point at the same path. that is, we are
3617 * dealing with a change.
3618 */
3619 if (one->sha1_valid && two->sha1_valid &&
3620 !hashcmp(one->sha1, two->sha1) &&
3621 !one->dirty_submodule && !two->dirty_submodule)
3622 return 1; /* no change */
3623 if (!one->sha1_valid && !two->sha1_valid)
3624 return 1; /* both look at the same file on the filesystem. */
3625 return 0;
3626 }
3627
3628 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3629 {
3630 if (diff_unmodified_pair(p))
3631 return;
3632
3633 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3634 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3635 return; /* no tree diffs in patch format */
3636
3637 run_diff(p, o);
3638 }
3639
3640 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3641 struct diffstat_t *diffstat)
3642 {
3643 if (diff_unmodified_pair(p))
3644 return;
3645
3646 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3647 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3648 return; /* no useful stat for tree diffs */
3649
3650 run_diffstat(p, o, diffstat);
3651 }
3652
3653 static void diff_flush_checkdiff(struct diff_filepair *p,
3654 struct diff_options *o)
3655 {
3656 if (diff_unmodified_pair(p))
3657 return;
3658
3659 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3660 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3661 return; /* nothing to check in tree diffs */
3662
3663 run_checkdiff(p, o);
3664 }
3665
3666 int diff_queue_is_empty(void)
3667 {
3668 struct diff_queue_struct *q = &diff_queued_diff;
3669 int i;
3670 for (i = 0; i < q->nr; i++)
3671 if (!diff_unmodified_pair(q->queue[i]))
3672 return 0;
3673 return 1;
3674 }
3675
3676 #if DIFF_DEBUG
3677 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3678 {
3679 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3680 x, one ? one : "",
3681 s->path,
3682 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3683 s->mode,
3684 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3685 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3686 x, one ? one : "",
3687 s->size, s->xfrm_flags);
3688 }
3689
3690 void diff_debug_filepair(const struct diff_filepair *p, int i)
3691 {
3692 diff_debug_filespec(p->one, i, "one");
3693 diff_debug_filespec(p->two, i, "two");
3694 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3695 p->score, p->status ? p->status : '?',
3696 p->one->rename_used, p->broken_pair);
3697 }
3698
3699 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3700 {
3701 int i;
3702 if (msg)
3703 fprintf(stderr, "%s\n", msg);
3704 fprintf(stderr, "q->nr = %d\n", q->nr);
3705 for (i = 0; i < q->nr; i++) {
3706 struct diff_filepair *p = q->queue[i];
3707 diff_debug_filepair(p, i);
3708 }
3709 }
3710 #endif
3711
3712 static void diff_resolve_rename_copy(void)
3713 {
3714 int i;
3715 struct diff_filepair *p;
3716 struct diff_queue_struct *q = &diff_queued_diff;
3717
3718 diff_debug_queue("resolve-rename-copy", q);
3719
3720 for (i = 0; i < q->nr; i++) {
3721 p = q->queue[i];
3722 p->status = 0; /* undecided */
3723 if (DIFF_PAIR_UNMERGED(p))
3724 p->status = DIFF_STATUS_UNMERGED;
3725 else if (!DIFF_FILE_VALID(p->one))
3726 p->status = DIFF_STATUS_ADDED;
3727 else if (!DIFF_FILE_VALID(p->two))
3728 p->status = DIFF_STATUS_DELETED;
3729 else if (DIFF_PAIR_TYPE_CHANGED(p))
3730 p->status = DIFF_STATUS_TYPE_CHANGED;
3731
3732 /* from this point on, we are dealing with a pair
3733 * whose both sides are valid and of the same type, i.e.
3734 * either in-place edit or rename/copy edit.
3735 */
3736 else if (DIFF_PAIR_RENAME(p)) {
3737 /*
3738 * A rename might have re-connected a broken
3739 * pair up, causing the pathnames to be the
3740 * same again. If so, that's not a rename at
3741 * all, just a modification..
3742 *
3743 * Otherwise, see if this source was used for
3744 * multiple renames, in which case we decrement
3745 * the count, and call it a copy.
3746 */
3747 if (!strcmp(p->one->path, p->two->path))
3748 p->status = DIFF_STATUS_MODIFIED;
3749 else if (--p->one->rename_used > 0)
3750 p->status = DIFF_STATUS_COPIED;
3751 else
3752 p->status = DIFF_STATUS_RENAMED;
3753 }
3754 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3755 p->one->mode != p->two->mode ||
3756 p->one->dirty_submodule ||
3757 p->two->dirty_submodule ||
3758 is_null_sha1(p->one->sha1))
3759 p->status = DIFF_STATUS_MODIFIED;
3760 else {
3761 /* This is a "no-change" entry and should not
3762 * happen anymore, but prepare for broken callers.
3763 */
3764 error("feeding unmodified %s to diffcore",
3765 p->one->path);
3766 p->status = DIFF_STATUS_UNKNOWN;
3767 }
3768 }
3769 diff_debug_queue("resolve-rename-copy done", q);
3770 }
3771
3772 static int check_pair_status(struct diff_filepair *p)
3773 {
3774 switch (p->status) {
3775 case DIFF_STATUS_UNKNOWN:
3776 return 0;
3777 case 0:
3778 die("internal error in diff-resolve-rename-copy");
3779 default:
3780 return 1;
3781 }
3782 }
3783
3784 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3785 {
3786 int fmt = opt->output_format;
3787
3788 if (fmt & DIFF_FORMAT_CHECKDIFF)
3789 diff_flush_checkdiff(p, opt);
3790 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3791 diff_flush_raw(p, opt);
3792 else if (fmt & DIFF_FORMAT_NAME) {
3793 const char *name_a, *name_b;
3794 name_a = p->two->path;
3795 name_b = NULL;
3796 strip_prefix(opt->prefix_length, &name_a, &name_b);
3797 write_name_quoted(name_a, opt->file, opt->line_termination);
3798 }
3799 }
3800
3801 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3802 {
3803 if (fs->mode)
3804 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3805 else
3806 fprintf(file, " %s ", newdelete);
3807 write_name_quoted(fs->path, file, '\n');
3808 }
3809
3810
3811 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3812 const char *line_prefix)
3813 {
3814 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3815 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3816 p->two->mode, show_name ? ' ' : '\n');
3817 if (show_name) {
3818 write_name_quoted(p->two->path, file, '\n');
3819 }
3820 }
3821 }
3822
3823 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3824 const char *line_prefix)
3825 {
3826 char *names = pprint_rename(p->one->path, p->two->path);
3827
3828 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3829 free(names);
3830 show_mode_change(file, p, 0, line_prefix);
3831 }
3832
3833 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3834 {
3835 FILE *file = opt->file;
3836 char *line_prefix = "";
3837
3838 if (opt->output_prefix) {
3839 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3840 line_prefix = buf->buf;
3841 }
3842
3843 switch(p->status) {
3844 case DIFF_STATUS_DELETED:
3845 fputs(line_prefix, file);
3846 show_file_mode_name(file, "delete", p->one);
3847 break;
3848 case DIFF_STATUS_ADDED:
3849 fputs(line_prefix, file);
3850 show_file_mode_name(file, "create", p->two);
3851 break;
3852 case DIFF_STATUS_COPIED:
3853 fputs(line_prefix, file);
3854 show_rename_copy(file, "copy", p, line_prefix);
3855 break;
3856 case DIFF_STATUS_RENAMED:
3857 fputs(line_prefix, file);
3858 show_rename_copy(file, "rename", p, line_prefix);
3859 break;
3860 default:
3861 if (p->score) {
3862 fprintf(file, "%s rewrite ", line_prefix);
3863 write_name_quoted(p->two->path, file, ' ');
3864 fprintf(file, "(%d%%)\n", similarity_index(p));
3865 }
3866 show_mode_change(file, p, !p->score, line_prefix);
3867 break;
3868 }
3869 }
3870
3871 struct patch_id_t {
3872 git_SHA_CTX *ctx;
3873 int patchlen;
3874 };
3875
3876 static int remove_space(char *line, int len)
3877 {
3878 int i;
3879 char *dst = line;
3880 unsigned char c;
3881
3882 for (i = 0; i < len; i++)
3883 if (!isspace((c = line[i])))
3884 *dst++ = c;
3885
3886 return dst - line;
3887 }
3888
3889 static void patch_id_consume(void *priv, char *line, unsigned long len)
3890 {
3891 struct patch_id_t *data = priv;
3892 int new_len;
3893
3894 /* Ignore line numbers when computing the SHA1 of the patch */
3895 if (!prefixcmp(line, "@@ -"))
3896 return;
3897
3898 new_len = remove_space(line, len);
3899
3900 git_SHA1_Update(data->ctx, line, new_len);
3901 data->patchlen += new_len;
3902 }
3903
3904 /* returns 0 upon success, and writes result into sha1 */
3905 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3906 {
3907 struct diff_queue_struct *q = &diff_queued_diff;
3908 int i;
3909 git_SHA_CTX ctx;
3910 struct patch_id_t data;
3911 char buffer[PATH_MAX * 4 + 20];
3912
3913 git_SHA1_Init(&ctx);
3914 memset(&data, 0, sizeof(struct patch_id_t));
3915 data.ctx = &ctx;
3916
3917 for (i = 0; i < q->nr; i++) {
3918 xpparam_t xpp;
3919 xdemitconf_t xecfg;
3920 mmfile_t mf1, mf2;
3921 struct diff_filepair *p = q->queue[i];
3922 int len1, len2;
3923
3924 memset(&xpp, 0, sizeof(xpp));
3925 memset(&xecfg, 0, sizeof(xecfg));
3926 if (p->status == 0)
3927 return error("internal diff status error");
3928 if (p->status == DIFF_STATUS_UNKNOWN)
3929 continue;
3930 if (diff_unmodified_pair(p))
3931 continue;
3932 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3933 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3934 continue;
3935 if (DIFF_PAIR_UNMERGED(p))
3936 continue;
3937
3938 diff_fill_sha1_info(p->one);
3939 diff_fill_sha1_info(p->two);
3940 if (fill_mmfile(&mf1, p->one) < 0 ||
3941 fill_mmfile(&mf2, p->two) < 0)
3942 return error("unable to read files to diff");
3943
3944 len1 = remove_space(p->one->path, strlen(p->one->path));
3945 len2 = remove_space(p->two->path, strlen(p->two->path));
3946 if (p->one->mode == 0)
3947 len1 = snprintf(buffer, sizeof(buffer),
3948 "diff--gita/%.*sb/%.*s"
3949 "newfilemode%06o"
3950 "---/dev/null"
3951 "+++b/%.*s",
3952 len1, p->one->path,
3953 len2, p->two->path,
3954 p->two->mode,
3955 len2, p->two->path);
3956 else if (p->two->mode == 0)
3957 len1 = snprintf(buffer, sizeof(buffer),
3958 "diff--gita/%.*sb/%.*s"
3959 "deletedfilemode%06o"
3960 "---a/%.*s"
3961 "+++/dev/null",
3962 len1, p->one->path,
3963 len2, p->two->path,
3964 p->one->mode,
3965 len1, p->one->path);
3966 else
3967 len1 = snprintf(buffer, sizeof(buffer),
3968 "diff--gita/%.*sb/%.*s"
3969 "---a/%.*s"
3970 "+++b/%.*s",
3971 len1, p->one->path,
3972 len2, p->two->path,
3973 len1, p->one->path,
3974 len2, p->two->path);
3975 git_SHA1_Update(&ctx, buffer, len1);
3976
3977 if (diff_filespec_is_binary(p->one) ||
3978 diff_filespec_is_binary(p->two)) {
3979 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3980 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3981 continue;
3982 }
3983
3984 xpp.flags = 0;
3985 xecfg.ctxlen = 3;
3986 xecfg.flags = 0;
3987 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3988 &xpp, &xecfg);
3989 }
3990
3991 git_SHA1_Final(sha1, &ctx);
3992 return 0;
3993 }
3994
3995 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3996 {
3997 struct diff_queue_struct *q = &diff_queued_diff;
3998 int i;
3999 int result = diff_get_patch_id(options, sha1);
4000
4001 for (i = 0; i < q->nr; i++)
4002 diff_free_filepair(q->queue[i]);
4003
4004 free(q->queue);
4005 DIFF_QUEUE_CLEAR(q);
4006
4007 return result;
4008 }
4009
4010 static int is_summary_empty(const struct diff_queue_struct *q)
4011 {
4012 int i;
4013
4014 for (i = 0; i < q->nr; i++) {
4015 const struct diff_filepair *p = q->queue[i];
4016
4017 switch (p->status) {
4018 case DIFF_STATUS_DELETED:
4019 case DIFF_STATUS_ADDED:
4020 case DIFF_STATUS_COPIED:
4021 case DIFF_STATUS_RENAMED:
4022 return 0;
4023 default:
4024 if (p->score)
4025 return 0;
4026 if (p->one->mode && p->two->mode &&
4027 p->one->mode != p->two->mode)
4028 return 0;
4029 break;
4030 }
4031 }
4032 return 1;
4033 }
4034
4035 static const char rename_limit_warning[] =
4036 "inexact rename detection was skipped due to too many files.";
4037
4038 static const char degrade_cc_to_c_warning[] =
4039 "only found copies from modified paths due to too many files.";
4040
4041 static const char rename_limit_advice[] =
4042 "you may want to set your %s variable to at least "
4043 "%d and retry the command.";
4044
4045 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4046 {
4047 if (degraded_cc)
4048 warning(degrade_cc_to_c_warning);
4049 else if (needed)
4050 warning(rename_limit_warning);
4051 else
4052 return;
4053 if (0 < needed && needed < 32767)
4054 warning(rename_limit_advice, varname, needed);
4055 }
4056
4057 void diff_flush(struct diff_options *options)
4058 {
4059 struct diff_queue_struct *q = &diff_queued_diff;
4060 int i, output_format = options->output_format;
4061 int separator = 0;
4062
4063 /*
4064 * Order: raw, stat, summary, patch
4065 * or: name/name-status/checkdiff (other bits clear)
4066 */
4067 if (!q->nr)
4068 goto free_queue;
4069
4070 if (output_format & (DIFF_FORMAT_RAW |
4071 DIFF_FORMAT_NAME |
4072 DIFF_FORMAT_NAME_STATUS |
4073 DIFF_FORMAT_CHECKDIFF)) {
4074 for (i = 0; i < q->nr; i++) {
4075 struct diff_filepair *p = q->queue[i];
4076 if (check_pair_status(p))
4077 flush_one_pair(p, options);
4078 }
4079 separator++;
4080 }
4081
4082 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4083 struct diffstat_t diffstat;
4084
4085 memset(&diffstat, 0, sizeof(struct diffstat_t));
4086 for (i = 0; i < q->nr; i++) {
4087 struct diff_filepair *p = q->queue[i];
4088 if (check_pair_status(p))
4089 diff_flush_stat(p, options, &diffstat);
4090 }
4091 if (output_format & DIFF_FORMAT_NUMSTAT)
4092 show_numstat(&diffstat, options);
4093 if (output_format & DIFF_FORMAT_DIFFSTAT)
4094 show_stats(&diffstat, options);
4095 if (output_format & DIFF_FORMAT_SHORTSTAT)
4096 show_shortstats(&diffstat, options);
4097 free_diffstat_info(&diffstat);
4098 separator++;
4099 }
4100 if (output_format & DIFF_FORMAT_DIRSTAT)
4101 show_dirstat(options);
4102
4103 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4104 for (i = 0; i < q->nr; i++) {
4105 diff_summary(options, q->queue[i]);
4106 }
4107 separator++;
4108 }
4109
4110 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4111 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4112 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4113 /*
4114 * run diff_flush_patch for the exit status. setting
4115 * options->file to /dev/null should be safe, becaue we
4116 * aren't supposed to produce any output anyway.
4117 */
4118 if (options->close_file)
4119 fclose(options->file);
4120 options->file = fopen("/dev/null", "w");
4121 if (!options->file)
4122 die_errno("Could not open /dev/null");
4123 options->close_file = 1;
4124 for (i = 0; i < q->nr; i++) {
4125 struct diff_filepair *p = q->queue[i];
4126 if (check_pair_status(p))
4127 diff_flush_patch(p, options);
4128 if (options->found_changes)
4129 break;
4130 }
4131 }
4132
4133 if (output_format & DIFF_FORMAT_PATCH) {
4134 if (separator) {
4135 putc(options->line_termination, options->file);
4136 if (options->stat_sep) {
4137 /* attach patch instead of inline */
4138 fputs(options->stat_sep, options->file);
4139 }
4140 }
4141
4142 for (i = 0; i < q->nr; i++) {
4143 struct diff_filepair *p = q->queue[i];
4144 if (check_pair_status(p))
4145 diff_flush_patch(p, options);
4146 }
4147 }
4148
4149 if (output_format & DIFF_FORMAT_CALLBACK)
4150 options->format_callback(q, options, options->format_callback_data);
4151
4152 for (i = 0; i < q->nr; i++)
4153 diff_free_filepair(q->queue[i]);
4154 free_queue:
4155 free(q->queue);
4156 DIFF_QUEUE_CLEAR(q);
4157 if (options->close_file)
4158 fclose(options->file);
4159
4160 /*
4161 * Report the content-level differences with HAS_CHANGES;
4162 * diff_addremove/diff_change does not set the bit when
4163 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4164 */
4165 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4166 if (options->found_changes)
4167 DIFF_OPT_SET(options, HAS_CHANGES);
4168 else
4169 DIFF_OPT_CLR(options, HAS_CHANGES);
4170 }
4171 }
4172
4173 static void diffcore_apply_filter(const char *filter)
4174 {
4175 int i;
4176 struct diff_queue_struct *q = &diff_queued_diff;
4177 struct diff_queue_struct outq;
4178 DIFF_QUEUE_CLEAR(&outq);
4179
4180 if (!filter)
4181 return;
4182
4183 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4184 int found;
4185 for (i = found = 0; !found && i < q->nr; i++) {
4186 struct diff_filepair *p = q->queue[i];
4187 if (((p->status == DIFF_STATUS_MODIFIED) &&
4188 ((p->score &&
4189 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4190 (!p->score &&
4191 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4192 ((p->status != DIFF_STATUS_MODIFIED) &&
4193 strchr(filter, p->status)))
4194 found++;
4195 }
4196 if (found)
4197 return;
4198
4199 /* otherwise we will clear the whole queue
4200 * by copying the empty outq at the end of this
4201 * function, but first clear the current entries
4202 * in the queue.
4203 */
4204 for (i = 0; i < q->nr; i++)
4205 diff_free_filepair(q->queue[i]);
4206 }
4207 else {
4208 /* Only the matching ones */
4209 for (i = 0; i < q->nr; i++) {
4210 struct diff_filepair *p = q->queue[i];
4211
4212 if (((p->status == DIFF_STATUS_MODIFIED) &&
4213 ((p->score &&
4214 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4215 (!p->score &&
4216 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4217 ((p->status != DIFF_STATUS_MODIFIED) &&
4218 strchr(filter, p->status)))
4219 diff_q(&outq, p);
4220 else
4221 diff_free_filepair(p);
4222 }
4223 }
4224 free(q->queue);
4225 *q = outq;
4226 }
4227
4228 /* Check whether two filespecs with the same mode and size are identical */
4229 static int diff_filespec_is_identical(struct diff_filespec *one,
4230 struct diff_filespec *two)
4231 {
4232 if (S_ISGITLINK(one->mode))
4233 return 0;
4234 if (diff_populate_filespec(one, 0))
4235 return 0;
4236 if (diff_populate_filespec(two, 0))
4237 return 0;
4238 return !memcmp(one->data, two->data, one->size);
4239 }
4240
4241 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4242 {
4243 int i;
4244 struct diff_queue_struct *q = &diff_queued_diff;
4245 struct diff_queue_struct outq;
4246 DIFF_QUEUE_CLEAR(&outq);
4247
4248 for (i = 0; i < q->nr; i++) {
4249 struct diff_filepair *p = q->queue[i];
4250
4251 /*
4252 * 1. Entries that come from stat info dirtiness
4253 * always have both sides (iow, not create/delete),
4254 * one side of the object name is unknown, with
4255 * the same mode and size. Keep the ones that
4256 * do not match these criteria. They have real
4257 * differences.
4258 *
4259 * 2. At this point, the file is known to be modified,
4260 * with the same mode and size, and the object
4261 * name of one side is unknown. Need to inspect
4262 * the identical contents.
4263 */
4264 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4265 !DIFF_FILE_VALID(p->two) ||
4266 (p->one->sha1_valid && p->two->sha1_valid) ||
4267 (p->one->mode != p->two->mode) ||
4268 diff_populate_filespec(p->one, 1) ||
4269 diff_populate_filespec(p->two, 1) ||
4270 (p->one->size != p->two->size) ||
4271 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4272 diff_q(&outq, p);
4273 else {
4274 /*
4275 * The caller can subtract 1 from skip_stat_unmatch
4276 * to determine how many paths were dirty only
4277 * due to stat info mismatch.
4278 */
4279 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4280 diffopt->skip_stat_unmatch++;
4281 diff_free_filepair(p);
4282 }
4283 }
4284 free(q->queue);
4285 *q = outq;
4286 }
4287
4288 static int diffnamecmp(const void *a_, const void *b_)
4289 {
4290 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4291 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4292 const char *name_a, *name_b;
4293
4294 name_a = a->one ? a->one->path : a->two->path;
4295 name_b = b->one ? b->one->path : b->two->path;
4296 return strcmp(name_a, name_b);
4297 }
4298
4299 void diffcore_fix_diff_index(struct diff_options *options)
4300 {
4301 struct diff_queue_struct *q = &diff_queued_diff;
4302 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4303 }
4304
4305 void diffcore_std(struct diff_options *options)
4306 {
4307 if (options->skip_stat_unmatch)
4308 diffcore_skip_stat_unmatch(options);
4309 if (!options->found_follow) {
4310 /* See try_to_follow_renames() in tree-diff.c */
4311 if (options->break_opt != -1)
4312 diffcore_break(options->break_opt);
4313 if (options->detect_rename)
4314 diffcore_rename(options);
4315 if (options->break_opt != -1)
4316 diffcore_merge_broken();
4317 }
4318 if (options->pickaxe)
4319 diffcore_pickaxe(options);
4320 if (options->orderfile)
4321 diffcore_order(options->orderfile);
4322 if (!options->found_follow)
4323 /* See try_to_follow_renames() in tree-diff.c */
4324 diff_resolve_rename_copy();
4325 diffcore_apply_filter(options->filter);
4326
4327 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4328 DIFF_OPT_SET(options, HAS_CHANGES);
4329 else
4330 DIFF_OPT_CLR(options, HAS_CHANGES);
4331
4332 options->found_follow = 0;
4333 }
4334
4335 int diff_result_code(struct diff_options *opt, int status)
4336 {
4337 int result = 0;
4338
4339 diff_warn_rename_limit("diff.renamelimit",
4340 opt->needed_rename_limit,
4341 opt->degraded_cc_to_c);
4342 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4343 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4344 return status;
4345 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4346 DIFF_OPT_TST(opt, HAS_CHANGES))
4347 result |= 01;
4348 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4349 DIFF_OPT_TST(opt, CHECK_FAILED))
4350 result |= 02;
4351 return result;
4352 }
4353
4354 /*
4355 * Shall changes to this submodule be ignored?
4356 *
4357 * Submodule changes can be configured to be ignored separately for each path,
4358 * but that configuration can be overridden from the command line.
4359 */
4360 static int is_submodule_ignored(const char *path, struct diff_options *options)
4361 {
4362 int ignored = 0;
4363 unsigned orig_flags = options->flags;
4364 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4365 set_diffopt_flags_from_submodule_config(options, path);
4366 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4367 ignored = 1;
4368 options->flags = orig_flags;
4369 return ignored;
4370 }
4371
4372 void diff_addremove(struct diff_options *options,
4373 int addremove, unsigned mode,
4374 const unsigned char *sha1,
4375 const char *concatpath, unsigned dirty_submodule)
4376 {
4377 struct diff_filespec *one, *two;
4378
4379 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4380 return;
4381
4382 /* This may look odd, but it is a preparation for
4383 * feeding "there are unchanged files which should
4384 * not produce diffs, but when you are doing copy
4385 * detection you would need them, so here they are"
4386 * entries to the diff-core. They will be prefixed
4387 * with something like '=' or '*' (I haven't decided
4388 * which but should not make any difference).
4389 * Feeding the same new and old to diff_change()
4390 * also has the same effect.
4391 * Before the final output happens, they are pruned after
4392 * merged into rename/copy pairs as appropriate.
4393 */
4394 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4395 addremove = (addremove == '+' ? '-' :
4396 addremove == '-' ? '+' : addremove);
4397
4398 if (options->prefix &&
4399 strncmp(concatpath, options->prefix, options->prefix_length))
4400 return;
4401
4402 one = alloc_filespec(concatpath);
4403 two = alloc_filespec(concatpath);
4404
4405 if (addremove != '+')
4406 fill_filespec(one, sha1, mode);
4407 if (addremove != '-') {
4408 fill_filespec(two, sha1, mode);
4409 two->dirty_submodule = dirty_submodule;
4410 }
4411
4412 diff_queue(&diff_queued_diff, one, two);
4413 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4414 DIFF_OPT_SET(options, HAS_CHANGES);
4415 }
4416
4417 void diff_change(struct diff_options *options,
4418 unsigned old_mode, unsigned new_mode,
4419 const unsigned char *old_sha1,
4420 const unsigned char *new_sha1,
4421 const char *concatpath,
4422 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4423 {
4424 struct diff_filespec *one, *two;
4425
4426 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4427 is_submodule_ignored(concatpath, options))
4428 return;
4429
4430 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4431 unsigned tmp;
4432 const unsigned char *tmp_c;
4433 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4434 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4435 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4436 new_dirty_submodule = tmp;
4437 }
4438
4439 if (options->prefix &&
4440 strncmp(concatpath, options->prefix, options->prefix_length))
4441 return;
4442
4443 one = alloc_filespec(concatpath);
4444 two = alloc_filespec(concatpath);
4445 fill_filespec(one, old_sha1, old_mode);
4446 fill_filespec(two, new_sha1, new_mode);
4447 one->dirty_submodule = old_dirty_submodule;
4448 two->dirty_submodule = new_dirty_submodule;
4449
4450 diff_queue(&diff_queued_diff, one, two);
4451 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4452 DIFF_OPT_SET(options, HAS_CHANGES);
4453 }
4454
4455 void diff_unmerge(struct diff_options *options,
4456 const char *path,
4457 unsigned mode, const unsigned char *sha1)
4458 {
4459 struct diff_filespec *one, *two;
4460
4461 if (options->prefix &&
4462 strncmp(path, options->prefix, options->prefix_length))
4463 return;
4464
4465 one = alloc_filespec(path);
4466 two = alloc_filespec(path);
4467 fill_filespec(one, sha1, mode);
4468 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4469 }
4470
4471 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4472 size_t *outsize)
4473 {
4474 struct diff_tempfile *temp;
4475 const char *argv[3];
4476 const char **arg = argv;
4477 struct child_process child;
4478 struct strbuf buf = STRBUF_INIT;
4479 int err = 0;
4480
4481 temp = prepare_temp_file(spec->path, spec);
4482 *arg++ = pgm;
4483 *arg++ = temp->name;
4484 *arg = NULL;
4485
4486 memset(&child, 0, sizeof(child));
4487 child.use_shell = 1;
4488 child.argv = argv;
4489 child.out = -1;
4490 if (start_command(&child)) {
4491 remove_tempfile();
4492 return NULL;
4493 }
4494
4495 if (strbuf_read(&buf, child.out, 0) < 0)
4496 err = error("error reading from textconv command '%s'", pgm);
4497 close(child.out);
4498
4499 if (finish_command(&child) || err) {
4500 strbuf_release(&buf);
4501 remove_tempfile();
4502 return NULL;
4503 }
4504 remove_tempfile();
4505
4506 return strbuf_detach(&buf, outsize);
4507 }
4508
4509 size_t fill_textconv(struct userdiff_driver *driver,
4510 struct diff_filespec *df,
4511 char **outbuf)
4512 {
4513 size_t size;
4514
4515 if (!driver || !driver->textconv) {
4516 if (!DIFF_FILE_VALID(df)) {
4517 *outbuf = "";
4518 return 0;
4519 }
4520 if (diff_populate_filespec(df, 0))
4521 die("unable to read files to diff");
4522 *outbuf = df->data;
4523 return df->size;
4524 }
4525
4526 if (driver->textconv_cache && df->sha1_valid) {
4527 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4528 &size);
4529 if (*outbuf)
4530 return size;
4531 }
4532
4533 *outbuf = run_textconv(driver->textconv, df, &size);
4534 if (!*outbuf)
4535 die("unable to read files to diff");
4536
4537 if (driver->textconv_cache && df->sha1_valid) {
4538 /* ignore errors, as we might be in a readonly repository */
4539 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4540 size);
4541 /*
4542 * we could save up changes and flush them all at the end,
4543 * but we would need an extra call after all diffing is done.
4544 * Since generating a cache entry is the slow path anyway,
4545 * this extra overhead probably isn't a big deal.
4546 */
4547 notes_cache_write(driver->textconv_cache);
4548 }
4549
4550 return size;
4551 }
Jonathan Nieder’s miscellaneous git branches