![[tokkee]](http://tokkee.org/images/avatar.png){kind=avatar}
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"
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
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;
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 };
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 }
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 }
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);
111 if (!strcmp(var, "diff.ignoresubmodules"))
112 handle_ignore_submodules_arg(&default_diff_options, value);
114 return git_diff_basic_config(var, value, cb);
115 }
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 }
124 switch (userdiff_config(var, value)) {
125 case 0: break;
126 case -1: return -1;
127 default: return 0;
128 }
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 }
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 }
148 if (!prefixcmp(var, "submodule."))
149 return parse_submodule_config_option(var, value);
151 return git_color_default_config(var, value, cb);
152 }
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;
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 }
172 static const char *external_diff(void)
173 {
174 static const char *external_diff_cmd = NULL;
175 static int done_preparing = 0;
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 }
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];
193 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
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 };
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 }
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;
243 mf->ptr = one->data;
244 mf->size = one->size;
245 return 0;
246 }
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 }
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;
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 }
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;
298 at = count_lines(mf2->ptr, mf2->size);
299 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
300 }
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;
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 }
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 }
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 }
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 }
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 }
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);
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 }
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;
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 @@ */
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));
411 /*
412 * trailing "\r\n"
413 */
414 for ( ; i < 3; i++)
415 if (line[len - i] == '\r' || line[len - i] == '\n')
416 len--;
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 }
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 }
434 strbuf_add(&msgbuf, line + len, org_len - len);
435 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
436 strbuf_release(&msgbuf);
437 }
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 }
447 static int remove_tempfile_installed;
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 }
459 static void remove_tempfile_on_signal(int signo)
460 {
461 remove_tempfile();
462 sigchain_pop(signo);
463 raise(signo);
464 }
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 }
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);
489 while (0 < size) {
490 int len;
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 }
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;
535 if (o && o->output_prefix) {
536 msgbuf = o->output_prefix(o, o->output_prefix_data);
537 line_prefix = msgbuf->buf;
538 }
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 }
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" : "";
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);
558 size_one = fill_textconv(textconv_one, one, &data_one);
559 size_two = fill_textconv(textconv_two, two, &data_two);
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;
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 }
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 };
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 }
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 };
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 };
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 };
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 };
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;
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 }
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 }
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 = "";
737 if (line[0] != '@' || parse_hunk_header(line, len,
738 &minus_first, &minus_len, &plus_first, &plus_len))
739 return;
741 assert(opt);
742 if (opt->output_prefix) {
743 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
744 line_prefix = msgbuf->buf;
745 }
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;
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;
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 }
786 diff_words->current_plus = plus_end;
787 diff_words->last_minus = minus_first;
788 }
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 }
806 /* find the next word */
807 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
808 (*begin)++;
809 if (*begin >= buffer->size)
810 return -1;
812 /* find the end of the word */
813 *end = *begin + 1;
814 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
815 (*end)++;
817 return 0;
818 }
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;
831 out->size = 0;
832 out->ptr = NULL;
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;
839 for (i = 0; i < buffer->text.size; i++) {
840 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
841 return;
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++;
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;
856 i = j - 1;
857 }
858 }
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;
868 struct diff_options *opt = diff_words->opt;
869 struct strbuf *msgbuf;
870 char *line_prefix = "";
872 assert(opt);
873 if (opt->output_prefix) {
874 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
875 line_prefix = msgbuf->buf;
876 }
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 }
889 diff_words->current_plus = diff_words->plus.text.ptr;
890 diff_words->last_minus = 0;
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 }
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 }
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 }
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 }
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;
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 }
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 }
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;
991 if (o && o->output_prefix) {
992 msgbuf = o->output_prefix(o, o->output_prefix_data);
993 line_prefix = msgbuf->buf;
994 }
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;
1003 if (ecbdata->label_path[0]) {
1004 const char *name_a_tab, *name_b_tab;
1006 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1007 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
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 }
1016 if (diff_suppress_blank_empty
1017 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1018 line[0] = '\n';
1019 len = 1;
1020 }
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 }
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 }
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 }
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 }
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);
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 }
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 }
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 }
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;
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 }
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 };
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 }
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];
1186 if (line[0] == '+')
1187 x->added++;
1188 else if (line[0] == '-')
1189 x->deleted++;
1190 }
1192 const char mime_boundary_leader[] = "------------";
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 }
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 }
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 }
1221 static void fill_print_name(struct diffstat_file *file)
1222 {
1223 char *pname;
1225 if (file->print_name)
1226 return;
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 }
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;
1248 const char *reset, *add_c, *del_c;
1249 const char *line_prefix = "";
1250 struct strbuf *msg = NULL;
1252 if (data->nr == 0)
1253 return;
1255 if (options->output_prefix) {
1256 msg = options->output_prefix(options, options->output_prefix_data);
1257 line_prefix = msg->buf;
1258 }
1260 width = options->stat_width ? options->stat_width : 80;
1261 name_width = options->stat_name_width ? options->stat_name_width : 50;
1263 /* Sanity: give at least 5 columns to the graph,
1264 * but leave at least 10 columns for the name.
1265 */
1266 if (width < 25)
1267 width = 25;
1268 if (name_width < 10)
1269 name_width = 10;
1270 else if (width < name_width + 15)
1271 name_width = width - 15;
1273 /* Find the longest filename and max number of changes */
1274 reset = diff_get_color_opt(options, DIFF_RESET);
1275 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1276 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1278 for (i = 0; i < data->nr; i++) {
1279 struct diffstat_file *file = data->files[i];
1280 uintmax_t change = file->added + file->deleted;
1281 if (!data->files[i]->is_renamed &&
1282 (change == 0)) {
1283 continue;
1284 }
1285 fill_print_name(file);
1286 len = strlen(file->print_name);
1287 if (max_len < len)
1288 max_len = len;
1290 if (file->is_binary || file->is_unmerged)
1291 continue;
1292 if (max_change < change)
1293 max_change = change;
1294 }
1296 /* Compute the width of the graph part;
1297 * 10 is for one blank at the beginning of the line plus
1298 * " | count " between the name and the graph.
1299 *
1300 * From here on, name_width is the width of the name area,
1301 * and width is the width of the graph area.
1302 */
1303 name_width = (name_width < max_len) ? name_width : max_len;
1304 if (width < (name_width + 10) + max_change)
1305 width = width - (name_width + 10);
1306 else
1307 width = max_change;
1309 for (i = 0; i < data->nr; i++) {
1310 const char *prefix = "";
1311 char *name = data->files[i]->print_name;
1312 uintmax_t added = data->files[i]->added;
1313 uintmax_t deleted = data->files[i]->deleted;
1314 int name_len;
1316 if (!data->files[i]->is_renamed &&
1317 (added + deleted == 0)) {
1318 total_files--;
1319 continue;
1320 }
1321 /*
1322 * "scale" the filename
1323 */
1324 len = name_width;
1325 name_len = strlen(name);
1326 if (name_width < name_len) {
1327 char *slash;
1328 prefix = "...";
1329 len -= 3;
1330 name += name_len - len;
1331 slash = strchr(name, '/');
1332 if (slash)
1333 name = slash;
1334 }
1336 if (data->files[i]->is_binary) {
1337 fprintf(options->file, "%s", line_prefix);
1338 show_name(options->file, prefix, name, len);
1339 fprintf(options->file, " Bin ");
1340 fprintf(options->file, "%s%"PRIuMAX"%s",
1341 del_c, deleted, reset);
1342 fprintf(options->file, " -> ");
1343 fprintf(options->file, "%s%"PRIuMAX"%s",
1344 add_c, added, reset);
1345 fprintf(options->file, " bytes");
1346 fprintf(options->file, "\n");
1347 continue;
1348 }
1349 else if (data->files[i]->is_unmerged) {
1350 fprintf(options->file, "%s", line_prefix);
1351 show_name(options->file, prefix, name, len);
1352 fprintf(options->file, " Unmerged\n");
1353 continue;
1354 }
1356 /*
1357 * scale the add/delete
1358 */
1359 add = added;
1360 del = deleted;
1361 adds += add;
1362 dels += del;
1364 if (width <= max_change) {
1365 add = scale_linear(add, width, max_change);
1366 del = scale_linear(del, width, max_change);
1367 }
1368 fprintf(options->file, "%s", line_prefix);
1369 show_name(options->file, prefix, name, len);
1370 fprintf(options->file, "%5"PRIuMAX"%s", added + deleted,
1371 added + deleted ? " " : "");
1372 show_graph(options->file, '+', add, add_c, reset);
1373 show_graph(options->file, '-', del, del_c, reset);
1374 fprintf(options->file, "\n");
1375 }
1376 fprintf(options->file, "%s", line_prefix);
1377 fprintf(options->file,
1378 " %d files changed, %d insertions(+), %d deletions(-)\n",
1379 total_files, adds, dels);
1380 }
1382 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1383 {
1384 int i, adds = 0, dels = 0, total_files = data->nr;
1386 if (data->nr == 0)
1387 return;
1389 for (i = 0; i < data->nr; i++) {
1390 if (!data->files[i]->is_binary &&
1391 !data->files[i]->is_unmerged) {
1392 int added = data->files[i]->added;
1393 int deleted= data->files[i]->deleted;
1394 if (!data->files[i]->is_renamed &&
1395 (added + deleted == 0)) {
1396 total_files--;
1397 } else {
1398 adds += added;
1399 dels += deleted;
1400 }
1401 }
1402 }
1403 if (options->output_prefix) {
1404 struct strbuf *msg = NULL;
1405 msg = options->output_prefix(options,
1406 options->output_prefix_data);
1407 fprintf(options->file, "%s", msg->buf);
1408 }
1409 fprintf(options->file, " %d files changed, %d insertions(+), %d deletions(-)\n",
1410 total_files, adds, dels);
1411 }
1413 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1414 {
1415 int i;
1417 if (data->nr == 0)
1418 return;
1420 for (i = 0; i < data->nr; i++) {
1421 struct diffstat_file *file = data->files[i];
1423 if (options->output_prefix) {
1424 struct strbuf *msg = NULL;
1425 msg = options->output_prefix(options,
1426 options->output_prefix_data);
1427 fprintf(options->file, "%s", msg->buf);
1428 }
1430 if (file->is_binary)
1431 fprintf(options->file, "-\t-\t");
1432 else
1433 fprintf(options->file,
1434 "%"PRIuMAX"\t%"PRIuMAX"\t",
1435 file->added, file->deleted);
1436 if (options->line_termination) {
1437 fill_print_name(file);
1438 if (!file->is_renamed)
1439 write_name_quoted(file->name, options->file,
1440 options->line_termination);
1441 else {
1442 fputs(file->print_name, options->file);
1443 putc(options->line_termination, options->file);
1444 }
1445 } else {
1446 if (file->is_renamed) {
1447 putc('\0', options->file);
1448 write_name_quoted(file->from_name, options->file, '\0');
1449 }
1450 write_name_quoted(file->name, options->file, '\0');
1451 }
1452 }
1453 }
1455 struct dirstat_file {
1456 const char *name;
1457 unsigned long changed;
1458 };
1460 struct dirstat_dir {
1461 struct dirstat_file *files;
1462 int alloc, nr, percent, cumulative;
1463 };
1465 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1466 unsigned long changed, const char *base, int baselen)
1467 {
1468 unsigned long this_dir = 0;
1469 unsigned int sources = 0;
1470 const char *line_prefix = "";
1471 struct strbuf *msg = NULL;
1473 if (opt->output_prefix) {
1474 msg = opt->output_prefix(opt, opt->output_prefix_data);
1475 line_prefix = msg->buf;
1476 }
1478 while (dir->nr) {
1479 struct dirstat_file *f = dir->files;
1480 int namelen = strlen(f->name);
1481 unsigned long this;
1482 char *slash;
1484 if (namelen < baselen)
1485 break;
1486 if (memcmp(f->name, base, baselen))
1487 break;
1488 slash = strchr(f->name + baselen, '/');
1489 if (slash) {
1490 int newbaselen = slash + 1 - f->name;
1491 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1492 sources++;
1493 } else {
1494 this = f->changed;
1495 dir->files++;
1496 dir->nr--;
1497 sources += 2;
1498 }
1499 this_dir += this;
1500 }
1502 /*
1503 * We don't report dirstat's for
1504 * - the top level
1505 * - or cases where everything came from a single directory
1506 * under this directory (sources == 1).
1507 */
1508 if (baselen && sources != 1) {
1509 int permille = this_dir * 1000 / changed;
1510 if (permille) {
1511 int percent = permille / 10;
1512 if (percent >= dir->percent) {
1513 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1514 percent, permille % 10, baselen, base);
1515 if (!dir->cumulative)
1516 return 0;
1517 }
1518 }
1519 }
1520 return this_dir;
1521 }
1523 static int dirstat_compare(const void *_a, const void *_b)
1524 {
1525 const struct dirstat_file *a = _a;
1526 const struct dirstat_file *b = _b;
1527 return strcmp(a->name, b->name);
1528 }
1530 static void show_dirstat(struct diff_options *options)
1531 {
1532 int i;
1533 unsigned long changed;
1534 struct dirstat_dir dir;
1535 struct diff_queue_struct *q = &diff_queued_diff;
1537 dir.files = NULL;
1538 dir.alloc = 0;
1539 dir.nr = 0;
1540 dir.percent = options->dirstat_percent;
1541 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1543 changed = 0;
1544 for (i = 0; i < q->nr; i++) {
1545 struct diff_filepair *p = q->queue[i];
1546 const char *name;
1547 unsigned long copied, added, damage;
1548 int content_changed;
1550 name = p->two->path ? p->two->path : p->one->path;
1552 if (p->one->sha1_valid && p->two->sha1_valid)
1553 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1554 else
1555 content_changed = 1;
1557 if (!content_changed) {
1558 /*
1559 * The SHA1 has not changed, so pre-/post-content is
1560 * identical. We can therefore skip looking at the
1561 * file contents altogether.
1562 */
1563 damage = 0;
1564 goto found_damage;
1565 }
1567 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1568 /*
1569 * In --dirstat-by-file mode, we don't really need to
1570 * look at the actual file contents at all.
1571 * The fact that the SHA1 changed is enough for us to
1572 * add this file to the list of results
1573 * (with each file contributing equal damage).
1574 */
1575 damage = 1;
1576 goto found_damage;
1577 }
1579 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1580 diff_populate_filespec(p->one, 0);
1581 diff_populate_filespec(p->two, 0);
1582 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1583 &copied, &added);
1584 diff_free_filespec_data(p->one);
1585 diff_free_filespec_data(p->two);
1586 } else if (DIFF_FILE_VALID(p->one)) {
1587 diff_populate_filespec(p->one, 1);
1588 copied = added = 0;
1589 diff_free_filespec_data(p->one);
1590 } else if (DIFF_FILE_VALID(p->two)) {
1591 diff_populate_filespec(p->two, 1);
1592 copied = 0;
1593 added = p->two->size;
1594 diff_free_filespec_data(p->two);
1595 } else
1596 continue;
1598 /*
1599 * Original minus copied is the removed material,
1600 * added is the new material. They are both damages
1601 * made to the preimage.
1602 * If the resulting damage is zero, we know that
1603 * diffcore_count_changes() considers the two entries to
1604 * be identical, but since content_changed is true, we
1605 * know that there must have been _some_ kind of change,
1606 * so we force all entries to have damage > 0.
1607 */
1608 damage = (p->one->size - copied) + added;
1609 if (!damage)
1610 damage = 1;
1612 found_damage:
1613 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1614 dir.files[dir.nr].name = name;
1615 dir.files[dir.nr].changed = damage;
1616 changed += damage;
1617 dir.nr++;
1618 }
1620 /* This can happen even with many files, if everything was renames */
1621 if (!changed)
1622 return;
1624 /* Show all directories with more than x% of the changes */
1625 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1626 gather_dirstat(options, &dir, changed, "", 0);
1627 }
1629 static void free_diffstat_info(struct diffstat_t *diffstat)
1630 {
1631 int i;
1632 for (i = 0; i < diffstat->nr; i++) {
1633 struct diffstat_file *f = diffstat->files[i];
1634 if (f->name != f->print_name)
1635 free(f->print_name);
1636 free(f->name);
1637 free(f->from_name);
1638 free(f);
1639 }
1640 free(diffstat->files);
1641 }
1643 struct checkdiff_t {
1644 const char *filename;
1645 int lineno;
1646 int conflict_marker_size;
1647 struct diff_options *o;
1648 unsigned ws_rule;
1649 unsigned status;
1650 };
1652 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1653 {
1654 char firstchar;
1655 int cnt;
1657 if (len < marker_size + 1)
1658 return 0;
1659 firstchar = line[0];
1660 switch (firstchar) {
1661 case '=': case '>': case '<': case '|':
1662 break;
1663 default:
1664 return 0;
1665 }
1666 for (cnt = 1; cnt < marker_size; cnt++)
1667 if (line[cnt] != firstchar)
1668 return 0;
1669 /* line[1] thru line[marker_size-1] are same as firstchar */
1670 if (len < marker_size + 1 || !isspace(line[marker_size]))
1671 return 0;
1672 return 1;
1673 }
1675 static void checkdiff_consume(void *priv, char *line, unsigned long len)
1676 {
1677 struct checkdiff_t *data = priv;
1678 int color_diff = DIFF_OPT_TST(data->o, COLOR_DIFF);
1679 int marker_size = data->conflict_marker_size;
1680 const char *ws = diff_get_color(color_diff, DIFF_WHITESPACE);
1681 const char *reset = diff_get_color(color_diff, DIFF_RESET);
1682 const char *set = diff_get_color(color_diff, DIFF_FILE_NEW);
1683 char *err;
1684 char *line_prefix = "";
1685 struct strbuf *msgbuf;
1687 assert(data->o);
1688 if (data->o->output_prefix) {
1689 msgbuf = data->o->output_prefix(data->o,
1690 data->o->output_prefix_data);
1691 line_prefix = msgbuf->buf;
1692 }
1694 if (line[0] == '+') {
1695 unsigned bad;
1696 data->lineno++;
1697 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
1698 data->status |= 1;
1699 fprintf(data->o->file,
1700 "%s%s:%d: leftover conflict marker\n",
1701 line_prefix, data->filename, data->lineno);
1702 }
1703 bad = ws_check(line + 1, len - 1, data->ws_rule);
1704 if (!bad)
1705 return;
1706 data->status |= bad;
1707 err = whitespace_error_string(bad);
1708 fprintf(data->o->file, "%s%s:%d: %s.\n",
1709 line_prefix, data->filename, data->lineno, err);
1710 free(err);
1711 emit_line(data->o, set, reset, line, 1);
1712 ws_check_emit(line + 1, len - 1, data->ws_rule,
1713 data->o->file, set, reset, ws);
1714 } else if (line[0] == ' ') {
1715 data->lineno++;
1716 } else if (line[0] == '@') {
1717 char *plus = strchr(line, '+');
1718 if (plus)
1719 data->lineno = strtol(plus, NULL, 10) - 1;
1720 else
1721 die("invalid diff");
1722 }
1723 }
1725 static unsigned char *deflate_it(char *data,
1726 unsigned long size,
1727 unsigned long *result_size)
1728 {
1729 int bound;
1730 unsigned char *deflated;
1731 z_stream stream;
1733 memset(&stream, 0, sizeof(stream));
1734 deflateInit(&stream, zlib_compression_level);
1735 bound = deflateBound(&stream, size);
1736 deflated = xmalloc(bound);
1737 stream.next_out = deflated;
1738 stream.avail_out = bound;
1740 stream.next_in = (unsigned char *)data;
1741 stream.avail_in = size;
1742 while (deflate(&stream, Z_FINISH) == Z_OK)
1743 ; /* nothing */
1744 deflateEnd(&stream);
1745 *result_size = stream.total_out;
1746 return deflated;
1747 }
1749 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1750 {
1751 void *cp;
1752 void *delta;
1753 void *deflated;
1754 void *data;
1755 unsigned long orig_size;
1756 unsigned long delta_size;
1757 unsigned long deflate_size;
1758 unsigned long data_size;
1760 /* We could do deflated delta, or we could do just deflated two,
1761 * whichever is smaller.
1762 */
1763 delta = NULL;
1764 deflated = deflate_it(two->ptr, two->size, &deflate_size);
1765 if (one->size && two->size) {
1766 delta = diff_delta(one->ptr, one->size,
1767 two->ptr, two->size,
1768 &delta_size, deflate_size);
1769 if (delta) {
1770 void *to_free = delta;
1771 orig_size = delta_size;
1772 delta = deflate_it(delta, delta_size, &delta_size);
1773 free(to_free);
1774 }
1775 }
1777 if (delta && delta_size < deflate_size) {
1778 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
1779 free(deflated);
1780 data = delta;
1781 data_size = delta_size;
1782 }
1783 else {
1784 fprintf(file, "%sliteral %lu\n", prefix, two->size);
1785 free(delta);
1786 data = deflated;
1787 data_size = deflate_size;
1788 }
1790 /* emit data encoded in base85 */
1791 cp = data;
1792 while (data_size) {
1793 int bytes = (52 < data_size) ? 52 : data_size;
1794 char line[70];
1795 data_size -= bytes;
1796 if (bytes <= 26)
1797 line[0] = bytes + 'A' - 1;
1798 else
1799 line[0] = bytes - 26 + 'a' - 1;
1800 encode_85(line + 1, cp, bytes);
1801 cp = (char *) cp + bytes;
1802 fprintf(file, "%s", prefix);
1803 fputs(line, file);
1804 fputc('\n', file);
1805 }
1806 fprintf(file, "%s\n", prefix);
1807 free(data);
1808 }
1810 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
1811 {
1812 fprintf(file, "%sGIT binary patch\n", prefix);
1813 emit_binary_diff_body(file, one, two, prefix);
1814 emit_binary_diff_body(file, two, one, prefix);
1815 }
1817 static void diff_filespec_load_driver(struct diff_filespec *one)
1818 {
1819 /* Use already-loaded driver */
1820 if (one->driver)
1821 return;
1823 if (S_ISREG(one->mode))
1824 one->driver = userdiff_find_by_path(one->path);
1826 /* Fallback to default settings */
1827 if (!one->driver)
1828 one->driver = userdiff_find_by_name("default");
1829 }
1831 int diff_filespec_is_binary(struct diff_filespec *one)
1832 {
1833 if (one->is_binary == -1) {
1834 diff_filespec_load_driver(one);
1835 if (one->driver->binary != -1)
1836 one->is_binary = one->driver->binary;
1837 else {
1838 if (!one->data && DIFF_FILE_VALID(one))
1839 diff_populate_filespec(one, 0);
1840 if (one->data)
1841 one->is_binary = buffer_is_binary(one->data,
1842 one->size);
1843 if (one->is_binary == -1)
1844 one->is_binary = 0;
1845 }
1846 }
1847 return one->is_binary;
1848 }
1850 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
1851 {
1852 diff_filespec_load_driver(one);
1853 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
1854 }
1856 static const char *userdiff_word_regex(struct diff_filespec *one)
1857 {
1858 diff_filespec_load_driver(one);
1859 return one->driver->word_regex;
1860 }
1862 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
1863 {
1864 if (!options->a_prefix)
1865 options->a_prefix = a;
1866 if (!options->b_prefix)
1867 options->b_prefix = b;
1868 }
1870 struct userdiff_driver *get_textconv(struct diff_filespec *one)
1871 {
1872 if (!DIFF_FILE_VALID(one))
1873 return NULL;
1875 diff_filespec_load_driver(one);
1876 if (!one->driver->textconv)
1877 return NULL;
1879 if (one->driver->textconv_want_cache && !one->driver->textconv_cache) {
1880 struct notes_cache *c = xmalloc(sizeof(*c));
1881 struct strbuf name = STRBUF_INIT;
1883 strbuf_addf(&name, "textconv/%s", one->driver->name);
1884 notes_cache_init(c, name.buf, one->driver->textconv);
1885 one->driver->textconv_cache = c;
1886 }
1888 return one->driver;
1889 }
1891 static void builtin_diff(const char *name_a,
1892 const char *name_b,
1893 struct diff_filespec *one,
1894 struct diff_filespec *two,
1895 const char *xfrm_msg,
1896 int must_show_header,
1897 struct diff_options *o,
1898 int complete_rewrite)
1899 {
1900 mmfile_t mf1, mf2;
1901 const char *lbl[2];
1902 char *a_one, *b_two;
1903 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
1904 const char *reset = diff_get_color_opt(o, DIFF_RESET);
1905 const char *a_prefix, *b_prefix;
1906 struct userdiff_driver *textconv_one = NULL;
1907 struct userdiff_driver *textconv_two = NULL;
1908 struct strbuf header = STRBUF_INIT;
1909 struct strbuf *msgbuf;
1910 char *line_prefix = "";
1912 if (o->output_prefix) {
1913 msgbuf = o->output_prefix(o, o->output_prefix_data);
1914 line_prefix = msgbuf->buf;
1915 }
1917 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
1918 (!one->mode || S_ISGITLINK(one->mode)) &&
1919 (!two->mode || S_ISGITLINK(two->mode))) {
1920 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
1921 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
1922 show_submodule_summary(o->file, one ? one->path : two->path,
1923 one->sha1, two->sha1, two->dirty_submodule,
1924 del, add, reset);
1925 return;
1926 }
1928 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
1929 textconv_one = get_textconv(one);
1930 textconv_two = get_textconv(two);
1931 }
1933 diff_set_mnemonic_prefix(o, "a/", "b/");
1934 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
1935 a_prefix = o->b_prefix;
1936 b_prefix = o->a_prefix;
1937 } else {
1938 a_prefix = o->a_prefix;
1939 b_prefix = o->b_prefix;
1940 }
1942 /* Never use a non-valid filename anywhere if at all possible */
1943 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
1944 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
1946 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
1947 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
1948 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
1949 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
1950 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
1951 if (lbl[0][0] == '/') {
1952 /* /dev/null */
1953 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
1954 if (xfrm_msg)
1955 strbuf_addstr(&header, xfrm_msg);
1956 must_show_header = 1;
1957 }
1958 else if (lbl[1][0] == '/') {
1959 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
1960 if (xfrm_msg)
1961 strbuf_addstr(&header, xfrm_msg);
1962 must_show_header = 1;
1963 }
1964 else {
1965 if (one->mode != two->mode) {
1966 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
1967 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
1968 must_show_header = 1;
1969 }
1970 if (xfrm_msg)
1971 strbuf_addstr(&header, xfrm_msg);
1973 /*
1974 * we do not run diff between different kind
1975 * of objects.
1976 */
1977 if ((one->mode ^ two->mode) & S_IFMT)
1978 goto free_ab_and_return;
1979 if (complete_rewrite &&
1980 (textconv_one || !diff_filespec_is_binary(one)) &&
1981 (textconv_two || !diff_filespec_is_binary(two))) {
1982 fprintf(o->file, "%s", header.buf);
1983 strbuf_reset(&header);
1984 emit_rewrite_diff(name_a, name_b, one, two,
1985 textconv_one, textconv_two, o);
1986 o->found_changes = 1;
1987 goto free_ab_and_return;
1988 }
1989 }
1991 if (o->irreversible_delete && lbl[1][0] == '/') {
1992 fprintf(o->file, "%s", header.buf);
1993 strbuf_reset(&header);
1994 goto free_ab_and_return;
1995 } else if (!DIFF_OPT_TST(o, TEXT) &&
1996 ( (!textconv_one && diff_filespec_is_binary(one)) ||
1997 (!textconv_two && diff_filespec_is_binary(two)) )) {
1998 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
1999 die("unable to read files to diff");
2000 /* Quite common confusing case */
2001 if (mf1.size == mf2.size &&
2002 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2003 if (must_show_header)
2004 fprintf(o->file, "%s", header.buf);
2005 goto free_ab_and_return;
2006 }
2007 fprintf(o->file, "%s", header.buf);
2008 strbuf_reset(&header);
2009 if (DIFF_OPT_TST(o, BINARY))
2010 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2011 else
2012 fprintf(o->file, "%sBinary files %s and %s differ\n",
2013 line_prefix, lbl[0], lbl[1]);
2014 o->found_changes = 1;
2015 } else {
2016 /* Crazy xdl interfaces.. */
2017 const char *diffopts = getenv("GIT_DIFF_OPTS");
2018 xpparam_t xpp;
2019 xdemitconf_t xecfg;
2020 struct emit_callback ecbdata;
2021 const struct userdiff_funcname *pe;
2023 if (!DIFF_XDL_TST(o, WHITESPACE_FLAGS) || must_show_header) {
2024 fprintf(o->file, "%s", header.buf);
2025 strbuf_reset(&header);
2026 }
2028 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2029 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2031 pe = diff_funcname_pattern(one);
2032 if (!pe)
2033 pe = diff_funcname_pattern(two);
2035 memset(&xpp, 0, sizeof(xpp));
2036 memset(&xecfg, 0, sizeof(xecfg));
2037 memset(&ecbdata, 0, sizeof(ecbdata));
2038 ecbdata.label_path = lbl;
2039 ecbdata.color_diff = DIFF_OPT_TST(o, COLOR_DIFF);
2040 ecbdata.found_changesp = &o->found_changes;
2041 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2042 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2043 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2044 ecbdata.opt = o;
2045 ecbdata.header = header.len ? &header : NULL;
2046 xpp.flags = o->xdl_opts;
2047 xecfg.ctxlen = o->context;
2048 xecfg.interhunkctxlen = o->interhunkcontext;
2049 xecfg.flags = XDL_EMIT_FUNCNAMES;
2050 if (pe)
2051 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2052 if (!diffopts)
2053 ;
2054 else if (!prefixcmp(diffopts, "--unified="))
2055 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2056 else if (!prefixcmp(diffopts, "-u"))
2057 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2058 if (o->word_diff) {
2059 int i;
2061 ecbdata.diff_words =
2062 xcalloc(1, sizeof(struct diff_words_data));
2063 ecbdata.diff_words->type = o->word_diff;
2064 ecbdata.diff_words->opt = o;
2065 if (!o->word_regex)
2066 o->word_regex = userdiff_word_regex(one);
2067 if (!o->word_regex)
2068 o->word_regex = userdiff_word_regex(two);
2069 if (!o->word_regex)
2070 o->word_regex = diff_word_regex_cfg;
2071 if (o->word_regex) {
2072 ecbdata.diff_words->word_regex = (regex_t *)
2073 xmalloc(sizeof(regex_t));
2074 if (regcomp(ecbdata.diff_words->word_regex,
2075 o->word_regex,
2076 REG_EXTENDED | REG_NEWLINE))
2077 die ("Invalid regular expression: %s",
2078 o->word_regex);
2079 }
2080 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
2081 if (o->word_diff == diff_words_styles[i].type) {
2082 ecbdata.diff_words->style =
2083 &diff_words_styles[i];
2084 break;
2085 }
2086 }
2087 if (DIFF_OPT_TST(o, COLOR_DIFF)) {
2088 struct diff_words_style *st = ecbdata.diff_words->style;
2089 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
2090 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
2091 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
2092 }
2093 }
2094 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2095 &xpp, &xecfg);
2096 if (o->word_diff)
2097 free_diff_words_data(&ecbdata);
2098 if (textconv_one)
2099 free(mf1.ptr);
2100 if (textconv_two)
2101 free(mf2.ptr);
2102 xdiff_clear_find_func(&xecfg);
2103 }
2105 free_ab_and_return:
2106 strbuf_release(&header);
2107 diff_free_filespec_data(one);
2108 diff_free_filespec_data(two);
2109 free(a_one);
2110 free(b_two);
2111 return;
2112 }
2114 static void builtin_diffstat(const char *name_a, const char *name_b,
2115 struct diff_filespec *one,
2116 struct diff_filespec *two,
2117 struct diffstat_t *diffstat,
2118 struct diff_options *o,
2119 int complete_rewrite)
2120 {
2121 mmfile_t mf1, mf2;
2122 struct diffstat_file *data;
2124 data = diffstat_add(diffstat, name_a, name_b);
2126 if (!one || !two) {
2127 data->is_unmerged = 1;
2128 return;
2129 }
2131 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2132 data->is_binary = 1;
2133 data->added = diff_filespec_size(two);
2134 data->deleted = diff_filespec_size(one);
2135 }
2137 else if (complete_rewrite) {
2138 diff_populate_filespec(one, 0);
2139 diff_populate_filespec(two, 0);
2140 data->deleted = count_lines(one->data, one->size);
2141 data->added = count_lines(two->data, two->size);
2142 }
2144 else {
2145 /* Crazy xdl interfaces.. */
2146 xpparam_t xpp;
2147 xdemitconf_t xecfg;
2149 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2150 die("unable to read files to diff");
2152 memset(&xpp, 0, sizeof(xpp));
2153 memset(&xecfg, 0, sizeof(xecfg));
2154 xpp.flags = o->xdl_opts;
2155 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2156 &xpp, &xecfg);
2157 }
2159 diff_free_filespec_data(one);
2160 diff_free_filespec_data(two);
2161 }
2163 static void builtin_checkdiff(const char *name_a, const char *name_b,
2164 const char *attr_path,
2165 struct diff_filespec *one,
2166 struct diff_filespec *two,
2167 struct diff_options *o)
2168 {
2169 mmfile_t mf1, mf2;
2170 struct checkdiff_t data;
2172 if (!two)
2173 return;
2175 memset(&data, 0, sizeof(data));
2176 data.filename = name_b ? name_b : name_a;
2177 data.lineno = 0;
2178 data.o = o;
2179 data.ws_rule = whitespace_rule(attr_path);
2180 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2182 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2183 die("unable to read files to diff");
2185 /*
2186 * All the other codepaths check both sides, but not checking
2187 * the "old" side here is deliberate. We are checking the newly
2188 * introduced changes, and as long as the "new" side is text, we
2189 * can and should check what it introduces.
2190 */
2191 if (diff_filespec_is_binary(two))
2192 goto free_and_return;
2193 else {
2194 /* Crazy xdl interfaces.. */
2195 xpparam_t xpp;
2196 xdemitconf_t xecfg;
2198 memset(&xpp, 0, sizeof(xpp));
2199 memset(&xecfg, 0, sizeof(xecfg));
2200 xecfg.ctxlen = 1; /* at least one context line */
2201 xpp.flags = 0;
2202 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2203 &xpp, &xecfg);
2205 if (data.ws_rule & WS_BLANK_AT_EOF) {
2206 struct emit_callback ecbdata;
2207 int blank_at_eof;
2209 ecbdata.ws_rule = data.ws_rule;
2210 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2211 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2213 if (blank_at_eof) {
2214 static char *err;
2215 if (!err)
2216 err = whitespace_error_string(WS_BLANK_AT_EOF);
2217 fprintf(o->file, "%s:%d: %s.\n",
2218 data.filename, blank_at_eof, err);
2219 data.status = 1; /* report errors */
2220 }
2221 }
2222 }
2223 free_and_return:
2224 diff_free_filespec_data(one);
2225 diff_free_filespec_data(two);
2226 if (data.status)
2227 DIFF_OPT_SET(o, CHECK_FAILED);
2228 }
2230 struct diff_filespec *alloc_filespec(const char *path)
2231 {
2232 int namelen = strlen(path);
2233 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2235 memset(spec, 0, sizeof(*spec));
2236 spec->path = (char *)(spec + 1);
2237 memcpy(spec->path, path, namelen+1);
2238 spec->count = 1;
2239 spec->is_binary = -1;
2240 return spec;
2241 }
2243 void free_filespec(struct diff_filespec *spec)
2244 {
2245 if (!--spec->count) {
2246 diff_free_filespec_data(spec);
2247 free(spec);
2248 }
2249 }
2251 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2252 unsigned short mode)
2253 {
2254 if (mode) {
2255 spec->mode = canon_mode(mode);
2256 hashcpy(spec->sha1, sha1);
2257 spec->sha1_valid = !is_null_sha1(sha1);
2258 }
2259 }
2261 /*
2262 * Given a name and sha1 pair, if the index tells us the file in
2263 * the work tree has that object contents, return true, so that
2264 * prepare_temp_file() does not have to inflate and extract.
2265 */
2266 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2267 {
2268 struct cache_entry *ce;
2269 struct stat st;
2270 int pos, len;
2272 /*
2273 * We do not read the cache ourselves here, because the
2274 * benchmark with my previous version that always reads cache
2275 * shows that it makes things worse for diff-tree comparing
2276 * two linux-2.6 kernel trees in an already checked out work
2277 * tree. This is because most diff-tree comparisons deal with
2278 * only a small number of files, while reading the cache is
2279 * expensive for a large project, and its cost outweighs the
2280 * savings we get by not inflating the object to a temporary
2281 * file. Practically, this code only helps when we are used
2282 * by diff-cache --cached, which does read the cache before
2283 * calling us.
2284 */
2285 if (!active_cache)
2286 return 0;
2288 /* We want to avoid the working directory if our caller
2289 * doesn't need the data in a normal file, this system
2290 * is rather slow with its stat/open/mmap/close syscalls,
2291 * and the object is contained in a pack file. The pack
2292 * is probably already open and will be faster to obtain
2293 * the data through than the working directory. Loose
2294 * objects however would tend to be slower as they need
2295 * to be individually opened and inflated.
2296 */
2297 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2298 return 0;
2300 len = strlen(name);
2301 pos = cache_name_pos(name, len);
2302 if (pos < 0)
2303 return 0;
2304 ce = active_cache[pos];
2306 /*
2307 * This is not the sha1 we are looking for, or
2308 * unreusable because it is not a regular file.
2309 */
2310 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2311 return 0;
2313 /*
2314 * If ce is marked as "assume unchanged", there is no
2315 * guarantee that work tree matches what we are looking for.
2316 */
2317 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2318 return 0;
2320 /*
2321 * If ce matches the file in the work tree, we can reuse it.
2322 */
2323 if (ce_uptodate(ce) ||
2324 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2325 return 1;
2327 return 0;
2328 }
2330 static int populate_from_stdin(struct diff_filespec *s)
2331 {
2332 struct strbuf buf = STRBUF_INIT;
2333 size_t size = 0;
2335 if (strbuf_read(&buf, 0, 0) < 0)
2336 return error("error while reading from stdin %s",
2337 strerror(errno));
2339 s->should_munmap = 0;
2340 s->data = strbuf_detach(&buf, &size);
2341 s->size = size;
2342 s->should_free = 1;
2343 return 0;
2344 }
2346 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2347 {
2348 int len;
2349 char *data = xmalloc(100), *dirty = "";
2351 /* Are we looking at the work tree? */
2352 if (s->dirty_submodule)
2353 dirty = "-dirty";
2355 len = snprintf(data, 100,
2356 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2357 s->data = data;
2358 s->size = len;
2359 s->should_free = 1;
2360 if (size_only) {
2361 s->data = NULL;
2362 free(data);
2363 }
2364 return 0;
2365 }
2367 /*
2368 * While doing rename detection and pickaxe operation, we may need to
2369 * grab the data for the blob (or file) for our own in-core comparison.
2370 * diff_filespec has data and size fields for this purpose.
2371 */
2372 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2373 {
2374 int err = 0;
2375 if (!DIFF_FILE_VALID(s))
2376 die("internal error: asking to populate invalid file.");
2377 if (S_ISDIR(s->mode))
2378 return -1;
2380 if (s->data)
2381 return 0;
2383 if (size_only && 0 < s->size)
2384 return 0;
2386 if (S_ISGITLINK(s->mode))
2387 return diff_populate_gitlink(s, size_only);
2389 if (!s->sha1_valid ||
2390 reuse_worktree_file(s->path, s->sha1, 0)) {
2391 struct strbuf buf = STRBUF_INIT;
2392 struct stat st;
2393 int fd;
2395 if (!strcmp(s->path, "-"))
2396 return populate_from_stdin(s);
2398 if (lstat(s->path, &st) < 0) {
2399 if (errno == ENOENT) {
2400 err_empty:
2401 err = -1;
2402 empty:
2403 s->data = (char *)"";
2404 s->size = 0;
2405 return err;
2406 }
2407 }
2408 s->size = xsize_t(st.st_size);
2409 if (!s->size)
2410 goto empty;
2411 if (S_ISLNK(st.st_mode)) {
2412 struct strbuf sb = STRBUF_INIT;
2414 if (strbuf_readlink(&sb, s->path, s->size))
2415 goto err_empty;
2416 s->size = sb.len;
2417 s->data = strbuf_detach(&sb, NULL);
2418 s->should_free = 1;
2419 return 0;
2420 }
2421 if (size_only)
2422 return 0;
2423 fd = open(s->path, O_RDONLY);
2424 if (fd < 0)
2425 goto err_empty;
2426 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2427 close(fd);
2428 s->should_munmap = 1;
2430 /*
2431 * Convert from working tree format to canonical git format
2432 */
2433 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2434 size_t size = 0;
2435 munmap(s->data, s->size);
2436 s->should_munmap = 0;
2437 s->data = strbuf_detach(&buf, &size);
2438 s->size = size;
2439 s->should_free = 1;
2440 }
2441 }
2442 else {
2443 enum object_type type;
2444 if (size_only) {
2445 type = sha1_object_info(s->sha1, &s->size);
2446 if (type < 0)
2447 die("unable to read %s", sha1_to_hex(s->sha1));
2448 } else {
2449 s->data = read_sha1_file(s->sha1, &type, &s->size);
2450 if (!s->data)
2451 die("unable to read %s", sha1_to_hex(s->sha1));
2452 s->should_free = 1;
2453 }
2454 }
2455 return 0;
2456 }
2458 void diff_free_filespec_blob(struct diff_filespec *s)
2459 {
2460 if (s->should_free)
2461 free(s->data);
2462 else if (s->should_munmap)
2463 munmap(s->data, s->size);
2465 if (s->should_free || s->should_munmap) {
2466 s->should_free = s->should_munmap = 0;
2467 s->data = NULL;
2468 }
2469 }
2471 void diff_free_filespec_data(struct diff_filespec *s)
2472 {
2473 diff_free_filespec_blob(s);
2474 free(s->cnt_data);
2475 s->cnt_data = NULL;
2476 }
2478 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2479 void *blob,
2480 unsigned long size,
2481 const unsigned char *sha1,
2482 int mode)
2483 {
2484 int fd;
2485 struct strbuf buf = STRBUF_INIT;
2486 struct strbuf template = STRBUF_INIT;
2487 char *path_dup = xstrdup(path);
2488 const char *base = basename(path_dup);
2490 /* Generate "XXXXXX_basename.ext" */
2491 strbuf_addstr(&template, "XXXXXX_");
2492 strbuf_addstr(&template, base);
2494 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2495 strlen(base) + 1);
2496 if (fd < 0)
2497 die_errno("unable to create temp-file");
2498 if (convert_to_working_tree(path,
2499 (const char *)blob, (size_t)size, &buf)) {
2500 blob = buf.buf;
2501 size = buf.len;
2502 }
2503 if (write_in_full(fd, blob, size) != size)
2504 die_errno("unable to write temp-file");
2505 close(fd);
2506 temp->name = temp->tmp_path;
2507 strcpy(temp->hex, sha1_to_hex(sha1));
2508 temp->hex[40] = 0;
2509 sprintf(temp->mode, "%06o", mode);
2510 strbuf_release(&buf);
2511 strbuf_release(&template);
2512 free(path_dup);
2513 }
2515 static struct diff_tempfile *prepare_temp_file(const char *name,
2516 struct diff_filespec *one)
2517 {
2518 struct diff_tempfile *temp = claim_diff_tempfile();
2520 if (!DIFF_FILE_VALID(one)) {
2521 not_a_valid_file:
2522 /* A '-' entry produces this for file-2, and
2523 * a '+' entry produces this for file-1.
2524 */
2525 temp->name = "/dev/null";
2526 strcpy(temp->hex, ".");
2527 strcpy(temp->mode, ".");
2528 return temp;
2529 }
2531 if (!remove_tempfile_installed) {
2532 atexit(remove_tempfile);
2533 sigchain_push_common(remove_tempfile_on_signal);
2534 remove_tempfile_installed = 1;
2535 }
2537 if (!one->sha1_valid ||
2538 reuse_worktree_file(name, one->sha1, 1)) {
2539 struct stat st;
2540 if (lstat(name, &st) < 0) {
2541 if (errno == ENOENT)
2542 goto not_a_valid_file;
2543 die_errno("stat(%s)", name);
2544 }
2545 if (S_ISLNK(st.st_mode)) {
2546 struct strbuf sb = STRBUF_INIT;
2547 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2548 die_errno("readlink(%s)", name);
2549 prep_temp_blob(name, temp, sb.buf, sb.len,
2550 (one->sha1_valid ?
2551 one->sha1 : null_sha1),
2552 (one->sha1_valid ?
2553 one->mode : S_IFLNK));
2554 strbuf_release(&sb);
2555 }
2556 else {
2557 /* we can borrow from the file in the work tree */
2558 temp->name = name;
2559 if (!one->sha1_valid)
2560 strcpy(temp->hex, sha1_to_hex(null_sha1));
2561 else
2562 strcpy(temp->hex, sha1_to_hex(one->sha1));
2563 /* Even though we may sometimes borrow the
2564 * contents from the work tree, we always want
2565 * one->mode. mode is trustworthy even when
2566 * !(one->sha1_valid), as long as
2567 * DIFF_FILE_VALID(one).
2568 */
2569 sprintf(temp->mode, "%06o", one->mode);
2570 }
2571 return temp;
2572 }
2573 else {
2574 if (diff_populate_filespec(one, 0))
2575 die("cannot read data blob for %s", one->path);
2576 prep_temp_blob(name, temp, one->data, one->size,
2577 one->sha1, one->mode);
2578 }
2579 return temp;
2580 }
2582 /* An external diff command takes:
2583 *
2584 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2585 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2586 *
2587 */
2588 static void run_external_diff(const char *pgm,
2589 const char *name,
2590 const char *other,
2591 struct diff_filespec *one,
2592 struct diff_filespec *two,
2593 const char *xfrm_msg,
2594 int complete_rewrite)
2595 {
2596 const char *spawn_arg[10];
2597 int retval;
2598 const char **arg = &spawn_arg[0];
2600 if (one && two) {
2601 struct diff_tempfile *temp_one, *temp_two;
2602 const char *othername = (other ? other : name);
2603 temp_one = prepare_temp_file(name, one);
2604 temp_two = prepare_temp_file(othername, two);
2605 *arg++ = pgm;
2606 *arg++ = name;
2607 *arg++ = temp_one->name;
2608 *arg++ = temp_one->hex;
2609 *arg++ = temp_one->mode;
2610 *arg++ = temp_two->name;
2611 *arg++ = temp_two->hex;
2612 *arg++ = temp_two->mode;
2613 if (other) {
2614 *arg++ = other;
2615 *arg++ = xfrm_msg;
2616 }
2617 } else {
2618 *arg++ = pgm;
2619 *arg++ = name;
2620 }
2621 *arg = NULL;
2622 fflush(NULL);
2623 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2624 remove_tempfile();
2625 if (retval) {
2626 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2627 exit(1);
2628 }
2629 }
2631 static int similarity_index(struct diff_filepair *p)
2632 {
2633 return p->score * 100 / MAX_SCORE;
2634 }
2636 static void fill_metainfo(struct strbuf *msg,
2637 const char *name,
2638 const char *other,
2639 struct diff_filespec *one,
2640 struct diff_filespec *two,
2641 struct diff_options *o,
2642 struct diff_filepair *p,
2643 int *must_show_header,
2644 int use_color)
2645 {
2646 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2647 const char *reset = diff_get_color(use_color, DIFF_RESET);
2648 struct strbuf *msgbuf;
2649 char *line_prefix = "";
2651 *must_show_header = 1;
2652 if (o->output_prefix) {
2653 msgbuf = o->output_prefix(o, o->output_prefix_data);
2654 line_prefix = msgbuf->buf;
2655 }
2656 strbuf_init(msg, PATH_MAX * 2 + 300);
2657 switch (p->status) {
2658 case DIFF_STATUS_COPIED:
2659 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2660 line_prefix, set, similarity_index(p));
2661 strbuf_addf(msg, "%s\n%s%scopy from ",
2662 reset, line_prefix, set);
2663 quote_c_style(name, msg, NULL, 0);
2664 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2665 quote_c_style(other, msg, NULL, 0);
2666 strbuf_addf(msg, "%s\n", reset);
2667 break;
2668 case DIFF_STATUS_RENAMED:
2669 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2670 line_prefix, set, similarity_index(p));
2671 strbuf_addf(msg, "%s\n%s%srename from ",
2672 reset, line_prefix, set);
2673 quote_c_style(name, msg, NULL, 0);
2674 strbuf_addf(msg, "%s\n%s%srename to ",
2675 reset, line_prefix, set);
2676 quote_c_style(other, msg, NULL, 0);
2677 strbuf_addf(msg, "%s\n", reset);
2678 break;
2679 case DIFF_STATUS_MODIFIED:
2680 if (p->score) {
2681 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2682 line_prefix,
2683 set, similarity_index(p), reset);
2684 break;
2685 }
2686 /* fallthru */
2687 default:
2688 *must_show_header = 0;
2689 }
2690 if (one && two && hashcmp(one->sha1, two->sha1)) {
2691 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2693 if (DIFF_OPT_TST(o, BINARY)) {
2694 mmfile_t mf;
2695 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2696 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2697 abbrev = 40;
2698 }
2699 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2700 find_unique_abbrev(one->sha1, abbrev));
2701 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2702 if (one->mode == two->mode)
2703 strbuf_addf(msg, " %06o", one->mode);
2704 strbuf_addf(msg, "%s\n", reset);
2705 }
2706 }
2708 static void run_diff_cmd(const char *pgm,
2709 const char *name,
2710 const char *other,
2711 const char *attr_path,
2712 struct diff_filespec *one,
2713 struct diff_filespec *two,
2714 struct strbuf *msg,
2715 struct diff_options *o,
2716 struct diff_filepair *p)
2717 {
2718 const char *xfrm_msg = NULL;
2719 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2720 int must_show_header = 0;
2722 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
2723 pgm = NULL;
2724 else {
2725 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
2726 if (drv && drv->external)
2727 pgm = drv->external;
2728 }
2730 if (msg) {
2731 /*
2732 * don't use colors when the header is intended for an
2733 * external diff driver
2734 */
2735 fill_metainfo(msg, name, other, one, two, o, p,
2736 &must_show_header,
2737 DIFF_OPT_TST(o, COLOR_DIFF) && !pgm);
2738 xfrm_msg = msg->len ? msg->buf : NULL;
2739 }
2741 if (pgm) {
2742 run_external_diff(pgm, name, other, one, two, xfrm_msg,
2743 complete_rewrite);
2744 return;
2745 }
2746 if (one && two)
2747 builtin_diff(name, other ? other : name,
2748 one, two, xfrm_msg, must_show_header,
2749 o, complete_rewrite);
2750 else
2751 fprintf(o->file, "* Unmerged path %s\n", name);
2752 }
2754 static void diff_fill_sha1_info(struct diff_filespec *one)
2755 {
2756 if (DIFF_FILE_VALID(one)) {
2757 if (!one->sha1_valid) {
2758 struct stat st;
2759 if (!strcmp(one->path, "-")) {
2760 hashcpy(one->sha1, null_sha1);
2761 return;
2762 }
2763 if (lstat(one->path, &st) < 0)
2764 die_errno("stat '%s'", one->path);
2765 if (index_path(one->sha1, one->path, &st, 0))
2766 die("cannot hash %s", one->path);
2767 }
2768 }
2769 else
2770 hashclr(one->sha1);
2771 }
2773 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
2774 {
2775 /* Strip the prefix but do not molest /dev/null and absolute paths */
2776 if (*namep && **namep != '/') {
2777 *namep += prefix_length;
2778 if (**namep == '/')
2779 ++*namep;
2780 }
2781 if (*otherp && **otherp != '/') {
2782 *otherp += prefix_length;
2783 if (**otherp == '/')
2784 ++*otherp;
2785 }
2786 }
2788 static void run_diff(struct diff_filepair *p, struct diff_options *o)
2789 {
2790 const char *pgm = external_diff();
2791 struct strbuf msg;
2792 struct diff_filespec *one = p->one;
2793 struct diff_filespec *two = p->two;
2794 const char *name;
2795 const char *other;
2796 const char *attr_path;
2798 name = p->one->path;
2799 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2800 attr_path = name;
2801 if (o->prefix_length)
2802 strip_prefix(o->prefix_length, &name, &other);
2804 if (DIFF_PAIR_UNMERGED(p)) {
2805 run_diff_cmd(pgm, name, NULL, attr_path,
2806 NULL, NULL, NULL, o, p);
2807 return;
2808 }
2810 diff_fill_sha1_info(one);
2811 diff_fill_sha1_info(two);
2813 if (!pgm &&
2814 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
2815 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
2816 /*
2817 * a filepair that changes between file and symlink
2818 * needs to be split into deletion and creation.
2819 */
2820 struct diff_filespec *null = alloc_filespec(two->path);
2821 run_diff_cmd(NULL, name, other, attr_path,
2822 one, null, &msg, o, p);
2823 free(null);
2824 strbuf_release(&msg);
2826 null = alloc_filespec(one->path);
2827 run_diff_cmd(NULL, name, other, attr_path,
2828 null, two, &msg, o, p);
2829 free(null);
2830 }
2831 else
2832 run_diff_cmd(pgm, name, other, attr_path,
2833 one, two, &msg, o, p);
2835 strbuf_release(&msg);
2836 }
2838 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
2839 struct diffstat_t *diffstat)
2840 {
2841 const char *name;
2842 const char *other;
2843 int complete_rewrite = 0;
2845 if (DIFF_PAIR_UNMERGED(p)) {
2846 /* unmerged */
2847 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
2848 return;
2849 }
2851 name = p->one->path;
2852 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2854 if (o->prefix_length)
2855 strip_prefix(o->prefix_length, &name, &other);
2857 diff_fill_sha1_info(p->one);
2858 diff_fill_sha1_info(p->two);
2860 if (p->status == DIFF_STATUS_MODIFIED && p->score)
2861 complete_rewrite = 1;
2862 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
2863 }
2865 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
2866 {
2867 const char *name;
2868 const char *other;
2869 const char *attr_path;
2871 if (DIFF_PAIR_UNMERGED(p)) {
2872 /* unmerged */
2873 return;
2874 }
2876 name = p->one->path;
2877 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
2878 attr_path = other ? other : name;
2880 if (o->prefix_length)
2881 strip_prefix(o->prefix_length, &name, &other);
2883 diff_fill_sha1_info(p->one);
2884 diff_fill_sha1_info(p->two);
2886 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
2887 }
2889 void diff_setup(struct diff_options *options)
2890 {
2891 memcpy(options, &default_diff_options, sizeof(*options));
2893 options->file = stdout;
2895 options->line_termination = '\n';
2896 options->break_opt = -1;
2897 options->rename_limit = -1;
2898 options->dirstat_percent = 3;
2899 options->context = 3;
2901 options->change = diff_change;
2902 options->add_remove = diff_addremove;
2903 if (diff_use_color_default > 0)
2904 DIFF_OPT_SET(options, COLOR_DIFF);
2905 options->detect_rename = diff_detect_rename_default;
2907 if (diff_no_prefix) {
2908 options->a_prefix = options->b_prefix = "";
2909 } else if (!diff_mnemonic_prefix) {
2910 options->a_prefix = "a/";
2911 options->b_prefix = "b/";
2912 }
2913 }
2915 int diff_setup_done(struct diff_options *options)
2916 {
2917 int count = 0;
2919 if (options->output_format & DIFF_FORMAT_NAME)
2920 count++;
2921 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
2922 count++;
2923 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
2924 count++;
2925 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
2926 count++;
2927 if (count > 1)
2928 die("--name-only, --name-status, --check and -s are mutually exclusive");
2930 /*
2931 * Most of the time we can say "there are changes"
2932 * only by checking if there are changed paths, but
2933 * --ignore-whitespace* options force us to look
2934 * inside contents.
2935 */
2937 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
2938 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
2939 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
2940 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
2941 else
2942 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
2944 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
2945 options->detect_rename = DIFF_DETECT_COPY;
2947 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
2948 options->prefix = NULL;
2949 if (options->prefix)
2950 options->prefix_length = strlen(options->prefix);
2951 else
2952 options->prefix_length = 0;
2954 if (options->output_format & (DIFF_FORMAT_NAME |
2955 DIFF_FORMAT_NAME_STATUS |
2956 DIFF_FORMAT_CHECKDIFF |
2957 DIFF_FORMAT_NO_OUTPUT))
2958 options->output_format &= ~(DIFF_FORMAT_RAW |
2959 DIFF_FORMAT_NUMSTAT |
2960 DIFF_FORMAT_DIFFSTAT |
2961 DIFF_FORMAT_SHORTSTAT |
2962 DIFF_FORMAT_DIRSTAT |
2963 DIFF_FORMAT_SUMMARY |
2964 DIFF_FORMAT_PATCH);
2966 /*
2967 * These cases always need recursive; we do not drop caller-supplied
2968 * recursive bits for other formats here.
2969 */
2970 if (options->output_format & (DIFF_FORMAT_PATCH |
2971 DIFF_FORMAT_NUMSTAT |
2972 DIFF_FORMAT_DIFFSTAT |
2973 DIFF_FORMAT_SHORTSTAT |
2974 DIFF_FORMAT_DIRSTAT |
2975 DIFF_FORMAT_SUMMARY |
2976 DIFF_FORMAT_CHECKDIFF))
2977 DIFF_OPT_SET(options, RECURSIVE);
2978 /*
2979 * Also pickaxe would not work very well if you do not say recursive
2980 */
2981 if (options->pickaxe)
2982 DIFF_OPT_SET(options, RECURSIVE);
2983 /*
2984 * When patches are generated, submodules diffed against the work tree
2985 * must be checked for dirtiness too so it can be shown in the output
2986 */
2987 if (options->output_format & DIFF_FORMAT_PATCH)
2988 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
2990 if (options->detect_rename && options->rename_limit < 0)
2991 options->rename_limit = diff_rename_limit_default;
2992 if (options->setup & DIFF_SETUP_USE_CACHE) {
2993 if (!active_cache)
2994 /* read-cache does not die even when it fails
2995 * so it is safe for us to do this here. Also
2996 * it does not smudge active_cache or active_nr
2997 * when it fails, so we do not have to worry about
2998 * cleaning it up ourselves either.
2999 */
3000 read_cache();
3001 }
3002 if (options->abbrev <= 0 || 40 < options->abbrev)
3003 options->abbrev = 40; /* full */
3005 /*
3006 * It does not make sense to show the first hit we happened
3007 * to have found. It does not make sense not to return with
3008 * exit code in such a case either.
3009 */
3010 if (DIFF_OPT_TST(options, QUICK)) {
3011 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3012 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3013 }
3015 return 0;
3016 }
3018 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3019 {
3020 char c, *eq;
3021 int len;
3023 if (*arg != '-')
3024 return 0;
3025 c = *++arg;
3026 if (!c)
3027 return 0;
3028 if (c == arg_short) {
3029 c = *++arg;
3030 if (!c)
3031 return 1;
3032 if (val && isdigit(c)) {
3033 char *end;
3034 int n = strtoul(arg, &end, 10);
3035 if (*end)
3036 return 0;
3037 *val = n;
3038 return 1;
3039 }
3040 return 0;
3041 }
3042 if (c != '-')
3043 return 0;
3044 arg++;
3045 eq = strchr(arg, '=');
3046 if (eq)
3047 len = eq - arg;
3048 else
3049 len = strlen(arg);
3050 if (!len || strncmp(arg, arg_long, len))
3051 return 0;
3052 if (eq) {
3053 int n;
3054 char *end;
3055 if (!isdigit(*++eq))
3056 return 0;
3057 n = strtoul(eq, &end, 10);
3058 if (*end)
3059 return 0;
3060 *val = n;
3061 }
3062 return 1;
3063 }
3065 static int diff_scoreopt_parse(const char *opt);
3067 static inline int short_opt(char opt, const char **argv,
3068 const char **optarg)
3069 {
3070 const char *arg = argv[0];
3071 if (arg[0] != '-' || arg[1] != opt)
3072 return 0;
3073 if (arg[2] != '\0') {
3074 *optarg = arg + 2;
3075 return 1;
3076 }
3077 if (!argv[1])
3078 die("Option '%c' requires a value", opt);
3079 *optarg = argv[1];
3080 return 2;
3081 }
3083 int parse_long_opt(const char *opt, const char **argv,
3084 const char **optarg)
3085 {
3086 const char *arg = argv[0];
3087 if (arg[0] != '-' || arg[1] != '-')
3088 return 0;
3089 arg += strlen("--");
3090 if (prefixcmp(arg, opt))
3091 return 0;
3092 arg += strlen(opt);
3093 if (*arg == '=') { /* sticked form: --option=value */
3094 *optarg = arg + 1;
3095 return 1;
3096 }
3097 if (*arg != '\0')
3098 return 0;
3099 /* separate form: --option value */
3100 if (!argv[1])
3101 die("Option '--%s' requires a value", opt);
3102 *optarg = argv[1];
3103 return 2;
3104 }
3106 static int stat_opt(struct diff_options *options, const char **av)
3107 {
3108 const char *arg = av[0];
3109 char *end;
3110 int width = options->stat_width;
3111 int name_width = options->stat_name_width;
3112 int argcount = 1;
3114 arg += strlen("--stat");
3115 end = (char *)arg;
3117 switch (*arg) {
3118 case '-':
3119 if (!prefixcmp(arg, "-width")) {
3120 arg += strlen("-width");
3121 if (*arg == '=')
3122 width = strtoul(arg + 1, &end, 10);
3123 else if (!*arg && !av[1])
3124 die("Option '--stat-width' requires a value");
3125 else if (!*arg) {
3126 width = strtoul(av[1], &end, 10);
3127 argcount = 2;
3128 }
3129 } else if (!prefixcmp(arg, "-name-width")) {
3130 arg += strlen("-name-width");
3131 if (*arg == '=')
3132 name_width = strtoul(arg + 1, &end, 10);
3133 else if (!*arg && !av[1])
3134 die("Option '--stat-name-width' requires a value");
3135 else if (!*arg) {
3136 name_width = strtoul(av[1], &end, 10);
3137 argcount = 2;
3138 }
3139 }
3140 break;
3141 case '=':
3142 width = strtoul(arg+1, &end, 10);
3143 if (*end == ',')
3144 name_width = strtoul(end+1, &end, 10);
3145 }
3147 /* Important! This checks all the error cases! */
3148 if (*end)
3149 return 0;
3150 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3151 options->stat_name_width = name_width;
3152 options->stat_width = width;
3153 return argcount;
3154 }
3156 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3157 {
3158 const char *arg = av[0];
3159 const char *optarg;
3160 int argcount;
3162 /* Output format options */
3163 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3164 options->output_format |= DIFF_FORMAT_PATCH;
3165 else if (opt_arg(arg, 'U', "unified", &options->context))
3166 options->output_format |= DIFF_FORMAT_PATCH;
3167 else if (!strcmp(arg, "--raw"))
3168 options->output_format |= DIFF_FORMAT_RAW;
3169 else if (!strcmp(arg, "--patch-with-raw"))
3170 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3171 else if (!strcmp(arg, "--numstat"))
3172 options->output_format |= DIFF_FORMAT_NUMSTAT;
3173 else if (!strcmp(arg, "--shortstat"))
3174 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3175 else if (opt_arg(arg, 'X', "dirstat", &options->dirstat_percent))
3176 options->output_format |= DIFF_FORMAT_DIRSTAT;
3177 else if (!strcmp(arg, "--cumulative")) {
3178 options->output_format |= DIFF_FORMAT_DIRSTAT;
3179 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
3180 } else if (opt_arg(arg, 0, "dirstat-by-file",
3181 &options->dirstat_percent)) {
3182 options->output_format |= DIFF_FORMAT_DIRSTAT;
3183 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
3184 }
3185 else if (!strcmp(arg, "--check"))
3186 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3187 else if (!strcmp(arg, "--summary"))
3188 options->output_format |= DIFF_FORMAT_SUMMARY;
3189 else if (!strcmp(arg, "--patch-with-stat"))
3190 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3191 else if (!strcmp(arg, "--name-only"))
3192 options->output_format |= DIFF_FORMAT_NAME;
3193 else if (!strcmp(arg, "--name-status"))
3194 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3195 else if (!strcmp(arg, "-s"))
3196 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3197 else if (!prefixcmp(arg, "--stat"))
3198 /* --stat, --stat-width, or --stat-name-width */
3199 return stat_opt(options, av);
3201 /* renames options */
3202 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3203 !strcmp(arg, "--break-rewrites")) {
3204 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3205 return error("invalid argument to -B: %s", arg+2);
3206 }
3207 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3208 !strcmp(arg, "--find-renames")) {
3209 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3210 return error("invalid argument to -M: %s", arg+2);
3211 options->detect_rename = DIFF_DETECT_RENAME;
3212 }
3213 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3214 options->irreversible_delete = 1;
3215 }
3216 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3217 !strcmp(arg, "--find-copies")) {
3218 if (options->detect_rename == DIFF_DETECT_COPY)
3219 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3220 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3221 return error("invalid argument to -C: %s", arg+2);
3222 options->detect_rename = DIFF_DETECT_COPY;
3223 }
3224 else if (!strcmp(arg, "--no-renames"))
3225 options->detect_rename = 0;
3226 else if (!strcmp(arg, "--relative"))
3227 DIFF_OPT_SET(options, RELATIVE_NAME);
3228 else if (!prefixcmp(arg, "--relative=")) {
3229 DIFF_OPT_SET(options, RELATIVE_NAME);
3230 options->prefix = arg + 11;
3231 }
3233 /* xdiff options */
3234 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3235 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3236 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3237 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3238 else if (!strcmp(arg, "--ignore-space-at-eol"))
3239 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3240 else if (!strcmp(arg, "--patience"))
3241 DIFF_XDL_SET(options, PATIENCE_DIFF);
3243 /* flags options */
3244 else if (!strcmp(arg, "--binary")) {
3245 options->output_format |= DIFF_FORMAT_PATCH;
3246 DIFF_OPT_SET(options, BINARY);
3247 }
3248 else if (!strcmp(arg, "--full-index"))
3249 DIFF_OPT_SET(options, FULL_INDEX);
3250 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3251 DIFF_OPT_SET(options, TEXT);
3252 else if (!strcmp(arg, "-R"))
3253 DIFF_OPT_SET(options, REVERSE_DIFF);
3254 else if (!strcmp(arg, "--find-copies-harder"))
3255 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3256 else if (!strcmp(arg, "--follow"))
3257 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3258 else if (!strcmp(arg, "--color"))
3259 DIFF_OPT_SET(options, COLOR_DIFF);
3260 else if (!prefixcmp(arg, "--color=")) {
3261 int value = git_config_colorbool(NULL, arg+8, -1);
3262 if (value == 0)
3263 DIFF_OPT_CLR(options, COLOR_DIFF);
3264 else if (value > 0)
3265 DIFF_OPT_SET(options, COLOR_DIFF);
3266 else
3267 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3268 }
3269 else if (!strcmp(arg, "--no-color"))
3270 DIFF_OPT_CLR(options, COLOR_DIFF);
3271 else if (!strcmp(arg, "--color-words")) {
3272 DIFF_OPT_SET(options, COLOR_DIFF);
3273 options->word_diff = DIFF_WORDS_COLOR;
3274 }
3275 else if (!prefixcmp(arg, "--color-words=")) {
3276 DIFF_OPT_SET(options, COLOR_DIFF);
3277 options->word_diff = DIFF_WORDS_COLOR;
3278 options->word_regex = arg + 14;
3279 }
3280 else if (!strcmp(arg, "--word-diff")) {
3281 if (options->word_diff == DIFF_WORDS_NONE)
3282 options->word_diff = DIFF_WORDS_PLAIN;
3283 }
3284 else if (!prefixcmp(arg, "--word-diff=")) {
3285 const char *type = arg + 12;
3286 if (!strcmp(type, "plain"))
3287 options->word_diff = DIFF_WORDS_PLAIN;
3288 else if (!strcmp(type, "color")) {
3289 DIFF_OPT_SET(options, COLOR_DIFF);
3290 options->word_diff = DIFF_WORDS_COLOR;
3291 }
3292 else if (!strcmp(type, "porcelain"))
3293 options->word_diff = DIFF_WORDS_PORCELAIN;
3294 else if (!strcmp(type, "none"))
3295 options->word_diff = DIFF_WORDS_NONE;
3296 else
3297 die("bad --word-diff argument: %s", type);
3298 }
3299 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3300 if (options->word_diff == DIFF_WORDS_NONE)
3301 options->word_diff = DIFF_WORDS_PLAIN;
3302 options->word_regex = optarg;
3303 return argcount;
3304 }
3305 else if (!strcmp(arg, "--exit-code"))
3306 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3307 else if (!strcmp(arg, "--quiet"))
3308 DIFF_OPT_SET(options, QUICK);
3309 else if (!strcmp(arg, "--ext-diff"))
3310 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3311 else if (!strcmp(arg, "--no-ext-diff"))
3312 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3313 else if (!strcmp(arg, "--textconv"))
3314 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3315 else if (!strcmp(arg, "--no-textconv"))
3316 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3317 else if (!strcmp(arg, "--ignore-submodules")) {
3318 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3319 handle_ignore_submodules_arg(options, "all");
3320 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3321 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3322 handle_ignore_submodules_arg(options, arg + 20);
3323 } else if (!strcmp(arg, "--submodule"))
3324 DIFF_OPT_SET(options, SUBMODULE_LOG);
3325 else if (!prefixcmp(arg, "--submodule=")) {
3326 if (!strcmp(arg + 12, "log"))
3327 DIFF_OPT_SET(options, SUBMODULE_LOG);
3328 }
3330 /* misc options */
3331 else if (!strcmp(arg, "-z"))
3332 options->line_termination = 0;
3333 else if ((argcount = short_opt('l', av, &optarg))) {
3334 options->rename_limit = strtoul(optarg, NULL, 10);
3335 return argcount;
3336 }
3337 else if ((argcount = short_opt('S', av, &optarg))) {
3338 options->pickaxe = optarg;
3339 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3340 return argcount;
3341 } else if ((argcount = short_opt('G', av, &optarg))) {
3342 options->pickaxe = optarg;
3343 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3344 return argcount;
3345 }
3346 else if (!strcmp(arg, "--pickaxe-all"))
3347 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3348 else if (!strcmp(arg, "--pickaxe-regex"))
3349 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3350 else if ((argcount = short_opt('O', av, &optarg))) {
3351 options->orderfile = optarg;
3352 return argcount;
3353 }
3354 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3355 options->filter = optarg;
3356 return argcount;
3357 }
3358 else if (!strcmp(arg, "--abbrev"))
3359 options->abbrev = DEFAULT_ABBREV;
3360 else if (!prefixcmp(arg, "--abbrev=")) {
3361 options->abbrev = strtoul(arg + 9, NULL, 10);
3362 if (options->abbrev < MINIMUM_ABBREV)
3363 options->abbrev = MINIMUM_ABBREV;
3364 else if (40 < options->abbrev)
3365 options->abbrev = 40;
3366 }
3367 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3368 options->a_prefix = optarg;
3369 return argcount;
3370 }
3371 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3372 options->b_prefix = optarg;
3373 return argcount;
3374 }
3375 else if (!strcmp(arg, "--no-prefix"))
3376 options->a_prefix = options->b_prefix = "";
3377 else if (opt_arg(arg, '\0', "inter-hunk-context",
3378 &options->interhunkcontext))
3379 ;
3380 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3381 options->file = fopen(optarg, "w");
3382 if (!options->file)
3383 die_errno("Could not open '%s'", optarg);
3384 options->close_file = 1;
3385 return argcount;
3386 } else
3387 return 0;
3388 return 1;
3389 }
3391 int parse_rename_score(const char **cp_p)
3392 {
3393 unsigned long num, scale;
3394 int ch, dot;
3395 const char *cp = *cp_p;
3397 num = 0;
3398 scale = 1;
3399 dot = 0;
3400 for (;;) {
3401 ch = *cp;
3402 if ( !dot && ch == '.' ) {
3403 scale = 1;
3404 dot = 1;
3405 } else if ( ch == '%' ) {
3406 scale = dot ? scale*100 : 100;
3407 cp++; /* % is always at the end */
3408 break;
3409 } else if ( ch >= '0' && ch <= '9' ) {
3410 if ( scale < 100000 ) {
3411 scale *= 10;
3412 num = (num*10) + (ch-'0');
3413 }
3414 } else {
3415 break;
3416 }
3417 cp++;
3418 }
3419 *cp_p = cp;
3421 /* user says num divided by scale and we say internally that
3422 * is MAX_SCORE * num / scale.
3423 */
3424 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3425 }
3427 static int diff_scoreopt_parse(const char *opt)
3428 {
3429 int opt1, opt2, cmd;
3431 if (*opt++ != '-')
3432 return -1;
3433 cmd = *opt++;
3434 if (cmd == '-') {
3435 /* convert the long-form arguments into short-form versions */
3436 if (!prefixcmp(opt, "break-rewrites")) {
3437 opt += strlen("break-rewrites");
3438 if (*opt == 0 || *opt++ == '=')
3439 cmd = 'B';
3440 } else if (!prefixcmp(opt, "find-copies")) {
3441 opt += strlen("find-copies");
3442 if (*opt == 0 || *opt++ == '=')
3443 cmd = 'C';
3444 } else if (!prefixcmp(opt, "find-renames")) {
3445 opt += strlen("find-renames");
3446 if (*opt == 0 || *opt++ == '=')
3447 cmd = 'M';
3448 }
3449 }
3450 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3451 return -1; /* that is not a -M, -C nor -B option */
3453 opt1 = parse_rename_score(&opt);
3454 if (cmd != 'B')
3455 opt2 = 0;
3456 else {
3457 if (*opt == 0)
3458 opt2 = 0;
3459 else if (*opt != '/')
3460 return -1; /* we expect -B80/99 or -B80 */
3461 else {
3462 opt++;
3463 opt2 = parse_rename_score(&opt);
3464 }
3465 }
3466 if (*opt != 0)
3467 return -1;
3468 return opt1 | (opt2 << 16);
3469 }
3471 struct diff_queue_struct diff_queued_diff;
3473 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3474 {
3475 if (queue->alloc <= queue->nr) {
3476 queue->alloc = alloc_nr(queue->alloc);
3477 queue->queue = xrealloc(queue->queue,
3478 sizeof(dp) * queue->alloc);
3479 }
3480 queue->queue[queue->nr++] = dp;
3481 }
3483 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3484 struct diff_filespec *one,
3485 struct diff_filespec *two)
3486 {
3487 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3488 dp->one = one;
3489 dp->two = two;
3490 if (queue)
3491 diff_q(queue, dp);
3492 return dp;
3493 }
3495 void diff_free_filepair(struct diff_filepair *p)
3496 {
3497 free_filespec(p->one);
3498 free_filespec(p->two);
3499 free(p);
3500 }
3502 /* This is different from find_unique_abbrev() in that
3503 * it stuffs the result with dots for alignment.
3504 */
3505 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3506 {
3507 int abblen;
3508 const char *abbrev;
3509 if (len == 40)
3510 return sha1_to_hex(sha1);
3512 abbrev = find_unique_abbrev(sha1, len);
3513 abblen = strlen(abbrev);
3514 if (abblen < 37) {
3515 static char hex[41];
3516 if (len < abblen && abblen <= len + 2)
3517 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3518 else
3519 sprintf(hex, "%s...", abbrev);
3520 return hex;
3521 }
3522 return sha1_to_hex(sha1);
3523 }
3525 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3526 {
3527 int line_termination = opt->line_termination;
3528 int inter_name_termination = line_termination ? '\t' : '\0';
3529 if (opt->output_prefix) {
3530 struct strbuf *msg = NULL;
3531 msg = opt->output_prefix(opt, opt->output_prefix_data);
3532 fprintf(opt->file, "%s", msg->buf);
3533 }
3535 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3536 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3537 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3538 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3539 }
3540 if (p->score) {
3541 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3542 inter_name_termination);
3543 } else {
3544 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3545 }
3547 if (p->status == DIFF_STATUS_COPIED ||
3548 p->status == DIFF_STATUS_RENAMED) {
3549 const char *name_a, *name_b;
3550 name_a = p->one->path;
3551 name_b = p->two->path;
3552 strip_prefix(opt->prefix_length, &name_a, &name_b);
3553 write_name_quoted(name_a, opt->file, inter_name_termination);
3554 write_name_quoted(name_b, opt->file, line_termination);
3555 } else {
3556 const char *name_a, *name_b;
3557 name_a = p->one->mode ? p->one->path : p->two->path;
3558 name_b = NULL;
3559 strip_prefix(opt->prefix_length, &name_a, &name_b);
3560 write_name_quoted(name_a, opt->file, line_termination);
3561 }
3562 }
3564 int diff_unmodified_pair(struct diff_filepair *p)
3565 {
3566 /* This function is written stricter than necessary to support
3567 * the currently implemented transformers, but the idea is to
3568 * let transformers to produce diff_filepairs any way they want,
3569 * and filter and clean them up here before producing the output.
3570 */
3571 struct diff_filespec *one = p->one, *two = p->two;
3573 if (DIFF_PAIR_UNMERGED(p))
3574 return 0; /* unmerged is interesting */
3576 /* deletion, addition, mode or type change
3577 * and rename are all interesting.
3578 */
3579 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3580 DIFF_PAIR_MODE_CHANGED(p) ||
3581 strcmp(one->path, two->path))
3582 return 0;
3584 /* both are valid and point at the same path. that is, we are
3585 * dealing with a change.
3586 */
3587 if (one->sha1_valid && two->sha1_valid &&
3588 !hashcmp(one->sha1, two->sha1) &&
3589 !one->dirty_submodule && !two->dirty_submodule)
3590 return 1; /* no change */
3591 if (!one->sha1_valid && !two->sha1_valid)
3592 return 1; /* both look at the same file on the filesystem. */
3593 return 0;
3594 }
3596 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3597 {
3598 if (diff_unmodified_pair(p))
3599 return;
3601 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3602 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3603 return; /* no tree diffs in patch format */
3605 run_diff(p, o);
3606 }
3608 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3609 struct diffstat_t *diffstat)
3610 {
3611 if (diff_unmodified_pair(p))
3612 return;
3614 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3615 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3616 return; /* no useful stat for tree diffs */
3618 run_diffstat(p, o, diffstat);
3619 }
3621 static void diff_flush_checkdiff(struct diff_filepair *p,
3622 struct diff_options *o)
3623 {
3624 if (diff_unmodified_pair(p))
3625 return;
3627 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3628 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3629 return; /* nothing to check in tree diffs */
3631 run_checkdiff(p, o);
3632 }
3634 int diff_queue_is_empty(void)
3635 {
3636 struct diff_queue_struct *q = &diff_queued_diff;
3637 int i;
3638 for (i = 0; i < q->nr; i++)
3639 if (!diff_unmodified_pair(q->queue[i]))
3640 return 0;
3641 return 1;
3642 }
3644 #if DIFF_DEBUG
3645 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3646 {
3647 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3648 x, one ? one : "",
3649 s->path,
3650 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3651 s->mode,
3652 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3653 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3654 x, one ? one : "",
3655 s->size, s->xfrm_flags);
3656 }
3658 void diff_debug_filepair(const struct diff_filepair *p, int i)
3659 {
3660 diff_debug_filespec(p->one, i, "one");
3661 diff_debug_filespec(p->two, i, "two");
3662 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3663 p->score, p->status ? p->status : '?',
3664 p->one->rename_used, p->broken_pair);
3665 }
3667 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
3668 {
3669 int i;
3670 if (msg)
3671 fprintf(stderr, "%s\n", msg);
3672 fprintf(stderr, "q->nr = %d\n", q->nr);
3673 for (i = 0; i < q->nr; i++) {
3674 struct diff_filepair *p = q->queue[i];
3675 diff_debug_filepair(p, i);
3676 }
3677 }
3678 #endif
3680 static void diff_resolve_rename_copy(void)
3681 {
3682 int i;
3683 struct diff_filepair *p;
3684 struct diff_queue_struct *q = &diff_queued_diff;
3686 diff_debug_queue("resolve-rename-copy", q);
3688 for (i = 0; i < q->nr; i++) {
3689 p = q->queue[i];
3690 p->status = 0; /* undecided */
3691 if (DIFF_PAIR_UNMERGED(p))
3692 p->status = DIFF_STATUS_UNMERGED;
3693 else if (!DIFF_FILE_VALID(p->one))
3694 p->status = DIFF_STATUS_ADDED;
3695 else if (!DIFF_FILE_VALID(p->two))
3696 p->status = DIFF_STATUS_DELETED;
3697 else if (DIFF_PAIR_TYPE_CHANGED(p))
3698 p->status = DIFF_STATUS_TYPE_CHANGED;
3700 /* from this point on, we are dealing with a pair
3701 * whose both sides are valid and of the same type, i.e.
3702 * either in-place edit or rename/copy edit.
3703 */
3704 else if (DIFF_PAIR_RENAME(p)) {
3705 /*
3706 * A rename might have re-connected a broken
3707 * pair up, causing the pathnames to be the
3708 * same again. If so, that's not a rename at
3709 * all, just a modification..
3710 *
3711 * Otherwise, see if this source was used for
3712 * multiple renames, in which case we decrement
3713 * the count, and call it a copy.
3714 */
3715 if (!strcmp(p->one->path, p->two->path))
3716 p->status = DIFF_STATUS_MODIFIED;
3717 else if (--p->one->rename_used > 0)
3718 p->status = DIFF_STATUS_COPIED;
3719 else
3720 p->status = DIFF_STATUS_RENAMED;
3721 }
3722 else if (hashcmp(p->one->sha1, p->two->sha1) ||
3723 p->one->mode != p->two->mode ||
3724 p->one->dirty_submodule ||
3725 p->two->dirty_submodule ||
3726 is_null_sha1(p->one->sha1))
3727 p->status = DIFF_STATUS_MODIFIED;
3728 else {
3729 /* This is a "no-change" entry and should not
3730 * happen anymore, but prepare for broken callers.
3731 */
3732 error("feeding unmodified %s to diffcore",
3733 p->one->path);
3734 p->status = DIFF_STATUS_UNKNOWN;
3735 }
3736 }
3737 diff_debug_queue("resolve-rename-copy done", q);
3738 }
3740 static int check_pair_status(struct diff_filepair *p)
3741 {
3742 switch (p->status) {
3743 case DIFF_STATUS_UNKNOWN:
3744 return 0;
3745 case 0:
3746 die("internal error in diff-resolve-rename-copy");
3747 default:
3748 return 1;
3749 }
3750 }
3752 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
3753 {
3754 int fmt = opt->output_format;
3756 if (fmt & DIFF_FORMAT_CHECKDIFF)
3757 diff_flush_checkdiff(p, opt);
3758 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
3759 diff_flush_raw(p, opt);
3760 else if (fmt & DIFF_FORMAT_NAME) {
3761 const char *name_a, *name_b;
3762 name_a = p->two->path;
3763 name_b = NULL;
3764 strip_prefix(opt->prefix_length, &name_a, &name_b);
3765 write_name_quoted(name_a, opt->file, opt->line_termination);
3766 }
3767 }
3769 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
3770 {
3771 if (fs->mode)
3772 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
3773 else
3774 fprintf(file, " %s ", newdelete);
3775 write_name_quoted(fs->path, file, '\n');
3776 }
3779 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
3780 const char *line_prefix)
3781 {
3782 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
3783 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
3784 p->two->mode, show_name ? ' ' : '\n');
3785 if (show_name) {
3786 write_name_quoted(p->two->path, file, '\n');
3787 }
3788 }
3789 }
3791 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
3792 const char *line_prefix)
3793 {
3794 char *names = pprint_rename(p->one->path, p->two->path);
3796 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
3797 free(names);
3798 show_mode_change(file, p, 0, line_prefix);
3799 }
3801 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
3802 {
3803 FILE *file = opt->file;
3804 char *line_prefix = "";
3806 if (opt->output_prefix) {
3807 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
3808 line_prefix = buf->buf;
3809 }
3811 switch(p->status) {
3812 case DIFF_STATUS_DELETED:
3813 fputs(line_prefix, file);
3814 show_file_mode_name(file, "delete", p->one);
3815 break;
3816 case DIFF_STATUS_ADDED:
3817 fputs(line_prefix, file);
3818 show_file_mode_name(file, "create", p->two);
3819 break;
3820 case DIFF_STATUS_COPIED:
3821 fputs(line_prefix, file);
3822 show_rename_copy(file, "copy", p, line_prefix);
3823 break;
3824 case DIFF_STATUS_RENAMED:
3825 fputs(line_prefix, file);
3826 show_rename_copy(file, "rename", p, line_prefix);
3827 break;
3828 default:
3829 if (p->score) {
3830 fprintf(file, "%s rewrite ", line_prefix);
3831 write_name_quoted(p->two->path, file, ' ');
3832 fprintf(file, "(%d%%)\n", similarity_index(p));
3833 }
3834 show_mode_change(file, p, !p->score, line_prefix);
3835 break;
3836 }
3837 }
3839 struct patch_id_t {
3840 git_SHA_CTX *ctx;
3841 int patchlen;
3842 };
3844 static int remove_space(char *line, int len)
3845 {
3846 int i;
3847 char *dst = line;
3848 unsigned char c;
3850 for (i = 0; i < len; i++)
3851 if (!isspace((c = line[i])))
3852 *dst++ = c;
3854 return dst - line;
3855 }
3857 static void patch_id_consume(void *priv, char *line, unsigned long len)
3858 {
3859 struct patch_id_t *data = priv;
3860 int new_len;
3862 /* Ignore line numbers when computing the SHA1 of the patch */
3863 if (!prefixcmp(line, "@@ -"))
3864 return;
3866 new_len = remove_space(line, len);
3868 git_SHA1_Update(data->ctx, line, new_len);
3869 data->patchlen += new_len;
3870 }
3872 /* returns 0 upon success, and writes result into sha1 */
3873 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
3874 {
3875 struct diff_queue_struct *q = &diff_queued_diff;
3876 int i;
3877 git_SHA_CTX ctx;
3878 struct patch_id_t data;
3879 char buffer[PATH_MAX * 4 + 20];
3881 git_SHA1_Init(&ctx);
3882 memset(&data, 0, sizeof(struct patch_id_t));
3883 data.ctx = &ctx;
3885 for (i = 0; i < q->nr; i++) {
3886 xpparam_t xpp;
3887 xdemitconf_t xecfg;
3888 mmfile_t mf1, mf2;
3889 struct diff_filepair *p = q->queue[i];
3890 int len1, len2;
3892 memset(&xpp, 0, sizeof(xpp));
3893 memset(&xecfg, 0, sizeof(xecfg));
3894 if (p->status == 0)
3895 return error("internal diff status error");
3896 if (p->status == DIFF_STATUS_UNKNOWN)
3897 continue;
3898 if (diff_unmodified_pair(p))
3899 continue;
3900 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3901 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3902 continue;
3903 if (DIFF_PAIR_UNMERGED(p))
3904 continue;
3906 diff_fill_sha1_info(p->one);
3907 diff_fill_sha1_info(p->two);
3908 if (fill_mmfile(&mf1, p->one) < 0 ||
3909 fill_mmfile(&mf2, p->two) < 0)
3910 return error("unable to read files to diff");
3912 len1 = remove_space(p->one->path, strlen(p->one->path));
3913 len2 = remove_space(p->two->path, strlen(p->two->path));
3914 if (p->one->mode == 0)
3915 len1 = snprintf(buffer, sizeof(buffer),
3916 "diff--gita/%.*sb/%.*s"
3917 "newfilemode%06o"
3918 "---/dev/null"
3919 "+++b/%.*s",
3920 len1, p->one->path,
3921 len2, p->two->path,
3922 p->two->mode,
3923 len2, p->two->path);
3924 else if (p->two->mode == 0)
3925 len1 = snprintf(buffer, sizeof(buffer),
3926 "diff--gita/%.*sb/%.*s"
3927 "deletedfilemode%06o"
3928 "---a/%.*s"
3929 "+++/dev/null",
3930 len1, p->one->path,
3931 len2, p->two->path,
3932 p->one->mode,
3933 len1, p->one->path);
3934 else
3935 len1 = snprintf(buffer, sizeof(buffer),
3936 "diff--gita/%.*sb/%.*s"
3937 "---a/%.*s"
3938 "+++b/%.*s",
3939 len1, p->one->path,
3940 len2, p->two->path,
3941 len1, p->one->path,
3942 len2, p->two->path);
3943 git_SHA1_Update(&ctx, buffer, len1);
3945 if (diff_filespec_is_binary(p->one) ||
3946 diff_filespec_is_binary(p->two)) {
3947 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
3948 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
3949 continue;
3950 }
3952 xpp.flags = 0;
3953 xecfg.ctxlen = 3;
3954 xecfg.flags = 0;
3955 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
3956 &xpp, &xecfg);
3957 }
3959 git_SHA1_Final(sha1, &ctx);
3960 return 0;
3961 }
3963 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
3964 {
3965 struct diff_queue_struct *q = &diff_queued_diff;
3966 int i;
3967 int result = diff_get_patch_id(options, sha1);
3969 for (i = 0; i < q->nr; i++)
3970 diff_free_filepair(q->queue[i]);
3972 free(q->queue);
3973 DIFF_QUEUE_CLEAR(q);
3975 return result;
3976 }
3978 static int is_summary_empty(const struct diff_queue_struct *q)
3979 {
3980 int i;
3982 for (i = 0; i < q->nr; i++) {
3983 const struct diff_filepair *p = q->queue[i];
3985 switch (p->status) {
3986 case DIFF_STATUS_DELETED:
3987 case DIFF_STATUS_ADDED:
3988 case DIFF_STATUS_COPIED:
3989 case DIFF_STATUS_RENAMED:
3990 return 0;
3991 default:
3992 if (p->score)
3993 return 0;
3994 if (p->one->mode && p->two->mode &&
3995 p->one->mode != p->two->mode)
3996 return 0;
3997 break;
3998 }
3999 }
4000 return 1;
4001 }
4003 static const char rename_limit_warning[] =
4004 "inexact rename detection was skipped due to too many files.";
4006 static const char degrade_cc_to_c_warning[] =
4007 "only found copies from modified paths due to too many files.";
4009 static const char rename_limit_advice[] =
4010 "you may want to set your %s variable to at least "
4011 "%d and retry the command.";
4013 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4014 {
4015 if (degraded_cc)
4016 warning(degrade_cc_to_c_warning);
4017 else if (needed)
4018 warning(rename_limit_warning);
4019 else
4020 return;
4021 if (0 < needed && needed < 32767)
4022 warning(rename_limit_advice, varname, needed);
4023 }
4025 void diff_flush(struct diff_options *options)
4026 {
4027 struct diff_queue_struct *q = &diff_queued_diff;
4028 int i, output_format = options->output_format;
4029 int separator = 0;
4031 /*
4032 * Order: raw, stat, summary, patch
4033 * or: name/name-status/checkdiff (other bits clear)
4034 */
4035 if (!q->nr)
4036 goto free_queue;
4038 if (output_format & (DIFF_FORMAT_RAW |
4039 DIFF_FORMAT_NAME |
4040 DIFF_FORMAT_NAME_STATUS |
4041 DIFF_FORMAT_CHECKDIFF)) {
4042 for (i = 0; i < q->nr; i++) {
4043 struct diff_filepair *p = q->queue[i];
4044 if (check_pair_status(p))
4045 flush_one_pair(p, options);
4046 }
4047 separator++;
4048 }
4050 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT)) {
4051 struct diffstat_t diffstat;
4053 memset(&diffstat, 0, sizeof(struct diffstat_t));
4054 for (i = 0; i < q->nr; i++) {
4055 struct diff_filepair *p = q->queue[i];
4056 if (check_pair_status(p))
4057 diff_flush_stat(p, options, &diffstat);
4058 }
4059 if (output_format & DIFF_FORMAT_NUMSTAT)
4060 show_numstat(&diffstat, options);
4061 if (output_format & DIFF_FORMAT_DIFFSTAT)
4062 show_stats(&diffstat, options);
4063 if (output_format & DIFF_FORMAT_SHORTSTAT)
4064 show_shortstats(&diffstat, options);
4065 free_diffstat_info(&diffstat);
4066 separator++;
4067 }
4068 if (output_format & DIFF_FORMAT_DIRSTAT)
4069 show_dirstat(options);
4071 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4072 for (i = 0; i < q->nr; i++) {
4073 diff_summary(options, q->queue[i]);
4074 }
4075 separator++;
4076 }
4078 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4079 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4080 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4081 /*
4082 * run diff_flush_patch for the exit status. setting
4083 * options->file to /dev/null should be safe, becaue we
4084 * aren't supposed to produce any output anyway.
4085 */
4086 if (options->close_file)
4087 fclose(options->file);
4088 options->file = fopen("/dev/null", "w");
4089 if (!options->file)
4090 die_errno("Could not open /dev/null");
4091 options->close_file = 1;
4092 for (i = 0; i < q->nr; i++) {
4093 struct diff_filepair *p = q->queue[i];
4094 if (check_pair_status(p))
4095 diff_flush_patch(p, options);
4096 if (options->found_changes)
4097 break;
4098 }
4099 }
4101 if (output_format & DIFF_FORMAT_PATCH) {
4102 if (separator) {
4103 putc(options->line_termination, options->file);
4104 if (options->stat_sep) {
4105 /* attach patch instead of inline */
4106 fputs(options->stat_sep, options->file);
4107 }
4108 }
4110 for (i = 0; i < q->nr; i++) {
4111 struct diff_filepair *p = q->queue[i];
4112 if (check_pair_status(p))
4113 diff_flush_patch(p, options);
4114 }
4115 }
4117 if (output_format & DIFF_FORMAT_CALLBACK)
4118 options->format_callback(q, options, options->format_callback_data);
4120 for (i = 0; i < q->nr; i++)
4121 diff_free_filepair(q->queue[i]);
4122 free_queue:
4123 free(q->queue);
4124 DIFF_QUEUE_CLEAR(q);
4125 if (options->close_file)
4126 fclose(options->file);
4128 /*
4129 * Report the content-level differences with HAS_CHANGES;
4130 * diff_addremove/diff_change does not set the bit when
4131 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4132 */
4133 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4134 if (options->found_changes)
4135 DIFF_OPT_SET(options, HAS_CHANGES);
4136 else
4137 DIFF_OPT_CLR(options, HAS_CHANGES);
4138 }
4139 }
4141 static void diffcore_apply_filter(const char *filter)
4142 {
4143 int i;
4144 struct diff_queue_struct *q = &diff_queued_diff;
4145 struct diff_queue_struct outq;
4146 DIFF_QUEUE_CLEAR(&outq);
4148 if (!filter)
4149 return;
4151 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4152 int found;
4153 for (i = found = 0; !found && i < q->nr; i++) {
4154 struct diff_filepair *p = q->queue[i];
4155 if (((p->status == DIFF_STATUS_MODIFIED) &&
4156 ((p->score &&
4157 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4158 (!p->score &&
4159 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4160 ((p->status != DIFF_STATUS_MODIFIED) &&
4161 strchr(filter, p->status)))
4162 found++;
4163 }
4164 if (found)
4165 return;
4167 /* otherwise we will clear the whole queue
4168 * by copying the empty outq at the end of this
4169 * function, but first clear the current entries
4170 * in the queue.
4171 */
4172 for (i = 0; i < q->nr; i++)
4173 diff_free_filepair(q->queue[i]);
4174 }
4175 else {
4176 /* Only the matching ones */
4177 for (i = 0; i < q->nr; i++) {
4178 struct diff_filepair *p = q->queue[i];
4180 if (((p->status == DIFF_STATUS_MODIFIED) &&
4181 ((p->score &&
4182 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4183 (!p->score &&
4184 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4185 ((p->status != DIFF_STATUS_MODIFIED) &&
4186 strchr(filter, p->status)))
4187 diff_q(&outq, p);
4188 else
4189 diff_free_filepair(p);
4190 }
4191 }
4192 free(q->queue);
4193 *q = outq;
4194 }
4196 /* Check whether two filespecs with the same mode and size are identical */
4197 static int diff_filespec_is_identical(struct diff_filespec *one,
4198 struct diff_filespec *two)
4199 {
4200 if (S_ISGITLINK(one->mode))
4201 return 0;
4202 if (diff_populate_filespec(one, 0))
4203 return 0;
4204 if (diff_populate_filespec(two, 0))
4205 return 0;
4206 return !memcmp(one->data, two->data, one->size);
4207 }
4209 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4210 {
4211 int i;
4212 struct diff_queue_struct *q = &diff_queued_diff;
4213 struct diff_queue_struct outq;
4214 DIFF_QUEUE_CLEAR(&outq);
4216 for (i = 0; i < q->nr; i++) {
4217 struct diff_filepair *p = q->queue[i];
4219 /*
4220 * 1. Entries that come from stat info dirtiness
4221 * always have both sides (iow, not create/delete),
4222 * one side of the object name is unknown, with
4223 * the same mode and size. Keep the ones that
4224 * do not match these criteria. They have real
4225 * differences.
4226 *
4227 * 2. At this point, the file is known to be modified,
4228 * with the same mode and size, and the object
4229 * name of one side is unknown. Need to inspect
4230 * the identical contents.
4231 */
4232 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4233 !DIFF_FILE_VALID(p->two) ||
4234 (p->one->sha1_valid && p->two->sha1_valid) ||
4235 (p->one->mode != p->two->mode) ||
4236 diff_populate_filespec(p->one, 1) ||
4237 diff_populate_filespec(p->two, 1) ||
4238 (p->one->size != p->two->size) ||
4239 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4240 diff_q(&outq, p);
4241 else {
4242 /*
4243 * The caller can subtract 1 from skip_stat_unmatch
4244 * to determine how many paths were dirty only
4245 * due to stat info mismatch.
4246 */
4247 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4248 diffopt->skip_stat_unmatch++;
4249 diff_free_filepair(p);
4250 }
4251 }
4252 free(q->queue);
4253 *q = outq;
4254 }
4256 static int diffnamecmp(const void *a_, const void *b_)
4257 {
4258 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4259 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4260 const char *name_a, *name_b;
4262 name_a = a->one ? a->one->path : a->two->path;
4263 name_b = b->one ? b->one->path : b->two->path;
4264 return strcmp(name_a, name_b);
4265 }
4267 void diffcore_fix_diff_index(struct diff_options *options)
4268 {
4269 struct diff_queue_struct *q = &diff_queued_diff;
4270 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4271 }
4273 void diffcore_std(struct diff_options *options)
4274 {
4275 if (options->skip_stat_unmatch)
4276 diffcore_skip_stat_unmatch(options);
4277 if (!options->found_follow) {
4278 /* See try_to_follow_renames() in tree-diff.c */
4279 if (options->break_opt != -1)
4280 diffcore_break(options->break_opt);
4281 if (options->detect_rename)
4282 diffcore_rename(options);
4283 if (options->break_opt != -1)
4284 diffcore_merge_broken();
4285 }
4286 if (options->pickaxe)
4287 diffcore_pickaxe(options);
4288 if (options->orderfile)
4289 diffcore_order(options->orderfile);
4290 if (!options->found_follow)
4291 /* See try_to_follow_renames() in tree-diff.c */
4292 diff_resolve_rename_copy();
4293 diffcore_apply_filter(options->filter);
4295 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4296 DIFF_OPT_SET(options, HAS_CHANGES);
4297 else
4298 DIFF_OPT_CLR(options, HAS_CHANGES);
4300 options->found_follow = 0;
4301 }
4303 int diff_result_code(struct diff_options *opt, int status)
4304 {
4305 int result = 0;
4307 diff_warn_rename_limit("diff.renamelimit",
4308 opt->needed_rename_limit,
4309 opt->degraded_cc_to_c);
4310 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4311 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4312 return status;
4313 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4314 DIFF_OPT_TST(opt, HAS_CHANGES))
4315 result |= 01;
4316 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4317 DIFF_OPT_TST(opt, CHECK_FAILED))
4318 result |= 02;
4319 return result;
4320 }
4322 /*
4323 * Shall changes to this submodule be ignored?
4324 *
4325 * Submodule changes can be configured to be ignored separately for each path,
4326 * but that configuration can be overridden from the command line.
4327 */
4328 static int is_submodule_ignored(const char *path, struct diff_options *options)
4329 {
4330 int ignored = 0;
4331 unsigned orig_flags = options->flags;
4332 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4333 set_diffopt_flags_from_submodule_config(options, path);
4334 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4335 ignored = 1;
4336 options->flags = orig_flags;
4337 return ignored;
4338 }
4340 void diff_addremove(struct diff_options *options,
4341 int addremove, unsigned mode,
4342 const unsigned char *sha1,
4343 const char *concatpath, unsigned dirty_submodule)
4344 {
4345 struct diff_filespec *one, *two;
4347 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4348 return;
4350 /* This may look odd, but it is a preparation for
4351 * feeding "there are unchanged files which should
4352 * not produce diffs, but when you are doing copy
4353 * detection you would need them, so here they are"
4354 * entries to the diff-core. They will be prefixed
4355 * with something like '=' or '*' (I haven't decided
4356 * which but should not make any difference).
4357 * Feeding the same new and old to diff_change()
4358 * also has the same effect.
4359 * Before the final output happens, they are pruned after
4360 * merged into rename/copy pairs as appropriate.
4361 */
4362 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4363 addremove = (addremove == '+' ? '-' :
4364 addremove == '-' ? '+' : addremove);
4366 if (options->prefix &&
4367 strncmp(concatpath, options->prefix, options->prefix_length))
4368 return;
4370 one = alloc_filespec(concatpath);
4371 two = alloc_filespec(concatpath);
4373 if (addremove != '+')
4374 fill_filespec(one, sha1, mode);
4375 if (addremove != '-') {
4376 fill_filespec(two, sha1, mode);
4377 two->dirty_submodule = dirty_submodule;
4378 }
4380 diff_queue(&diff_queued_diff, one, two);
4381 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4382 DIFF_OPT_SET(options, HAS_CHANGES);
4383 }
4385 void diff_change(struct diff_options *options,
4386 unsigned old_mode, unsigned new_mode,
4387 const unsigned char *old_sha1,
4388 const unsigned char *new_sha1,
4389 const char *concatpath,
4390 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4391 {
4392 struct diff_filespec *one, *two;
4394 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4395 is_submodule_ignored(concatpath, options))
4396 return;
4398 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4399 unsigned tmp;
4400 const unsigned char *tmp_c;
4401 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4402 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4403 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4404 new_dirty_submodule = tmp;
4405 }
4407 if (options->prefix &&
4408 strncmp(concatpath, options->prefix, options->prefix_length))
4409 return;
4411 one = alloc_filespec(concatpath);
4412 two = alloc_filespec(concatpath);
4413 fill_filespec(one, old_sha1, old_mode);
4414 fill_filespec(two, new_sha1, new_mode);
4415 one->dirty_submodule = old_dirty_submodule;
4416 two->dirty_submodule = new_dirty_submodule;
4418 diff_queue(&diff_queued_diff, one, two);
4419 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4420 DIFF_OPT_SET(options, HAS_CHANGES);
4421 }
4423 void diff_unmerge(struct diff_options *options,
4424 const char *path,
4425 unsigned mode, const unsigned char *sha1)
4426 {
4427 struct diff_filespec *one, *two;
4429 if (options->prefix &&
4430 strncmp(path, options->prefix, options->prefix_length))
4431 return;
4433 one = alloc_filespec(path);
4434 two = alloc_filespec(path);
4435 fill_filespec(one, sha1, mode);
4436 diff_queue(&diff_queued_diff, one, two)->is_unmerged = 1;
4437 }
4439 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4440 size_t *outsize)
4441 {
4442 struct diff_tempfile *temp;
4443 const char *argv[3];
4444 const char **arg = argv;
4445 struct child_process child;
4446 struct strbuf buf = STRBUF_INIT;
4447 int err = 0;
4449 temp = prepare_temp_file(spec->path, spec);
4450 *arg++ = pgm;
4451 *arg++ = temp->name;
4452 *arg = NULL;
4454 memset(&child, 0, sizeof(child));
4455 child.use_shell = 1;
4456 child.argv = argv;
4457 child.out = -1;
4458 if (start_command(&child)) {
4459 remove_tempfile();
4460 return NULL;
4461 }
4463 if (strbuf_read(&buf, child.out, 0) < 0)
4464 err = error("error reading from textconv command '%s'", pgm);
4465 close(child.out);
4467 if (finish_command(&child) || err) {
4468 strbuf_release(&buf);
4469 remove_tempfile();
4470 return NULL;
4471 }
4472 remove_tempfile();
4474 return strbuf_detach(&buf, outsize);
4475 }
4477 size_t fill_textconv(struct userdiff_driver *driver,
4478 struct diff_filespec *df,
4479 char **outbuf)
4480 {
4481 size_t size;
4483 if (!driver || !driver->textconv) {
4484 if (!DIFF_FILE_VALID(df)) {
4485 *outbuf = "";
4486 return 0;
4487 }
4488 if (diff_populate_filespec(df, 0))
4489 die("unable to read files to diff");
4490 *outbuf = df->data;
4491 return df->size;
4492 }
4494 if (driver->textconv_cache && df->sha1_valid) {
4495 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4496 &size);
4497 if (*outbuf)
4498 return size;
4499 }
4501 *outbuf = run_textconv(driver->textconv, df, &size);
4502 if (!*outbuf)
4503 die("unable to read files to diff");
4505 if (driver->textconv_cache && df->sha1_valid) {
4506 /* ignore errors, as we might be in a readonly repository */
4507 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4508 size);
4509 /*
4510 * we could save up changes and flush them all at the end,
4511 * but we would need an extra call after all diffing is done.
4512 * Since generating a cache entry is the slow path anyway,
4513 * this extra overhead probably isn't a big deal.
4514 */
4515 notes_cache_write(driver->textconv_cache);
4516 }
4518 return size;
4519 }