?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
diff --stat: use less columns for change counts
[git/jnareb-git.git] / diff.c
blobc316e0cbfbb788a6e0066069c3425b1f712e543b
1 /*
2 * Copyright (C) 2005 Junio C Hamano
3 */
4 #include "cache.h"
5 #include "quote.h"
6 #include "diff.h"
7 #include "diffcore.h"
8 #include "delta.h"
9 #include "xdiff-interface.h"
10 #include "color.h"
11 #include "attr.h"
12 #include "run-command.h"
13 #include "utf8.h"
14 #include "userdiff.h"
15 #include "sigchain.h"
16 #include "submodule.h"
17 #include "ll-merge.h"
18
19 #ifdef NO_FAST_WORKING_DIRECTORY
20 #define FAST_WORKING_DIRECTORY 0
21 #else
22 #define FAST_WORKING_DIRECTORY 1
23 #endif
24
25 static int diff_detect_rename_default;
26 static int diff_rename_limit_default = 400;
27 static int diff_suppress_blank_empty;
28 int diff_use_color_default = -1;
29 static const char *diff_word_regex_cfg;
30 static const char *external_diff_cmd_cfg;
31 int diff_auto_refresh_index = 1;
32 static int diff_mnemonic_prefix;
33 static int diff_no_prefix;
34 static int diff_stat_graph_width;
35 static int diff_dirstat_permille_default = 30;
36 static struct diff_options default_diff_options;
37
38 static char diff_colors[][COLOR_MAXLEN] = {
39 GIT_COLOR_RESET,
40 GIT_COLOR_NORMAL, /* PLAIN */
41 GIT_COLOR_BOLD, /* METAINFO */
42 GIT_COLOR_CYAN, /* FRAGINFO */
43 GIT_COLOR_RED, /* OLD */
44 GIT_COLOR_GREEN, /* NEW */
45 GIT_COLOR_YELLOW, /* COMMIT */
46 GIT_COLOR_BG_RED, /* WHITESPACE */
47 GIT_COLOR_NORMAL, /* FUNCINFO */
48 };
49
50 static int parse_diff_color_slot(const char *var, int ofs)
51 {
52 if (!strcasecmp(var+ofs, "plain"))
53 return DIFF_PLAIN;
54 if (!strcasecmp(var+ofs, "meta"))
55 return DIFF_METAINFO;
56 if (!strcasecmp(var+ofs, "frag"))
57 return DIFF_FRAGINFO;
58 if (!strcasecmp(var+ofs, "old"))
59 return DIFF_FILE_OLD;
60 if (!strcasecmp(var+ofs, "new"))
61 return DIFF_FILE_NEW;
62 if (!strcasecmp(var+ofs, "commit"))
63 return DIFF_COMMIT;
64 if (!strcasecmp(var+ofs, "whitespace"))
65 return DIFF_WHITESPACE;
66 if (!strcasecmp(var+ofs, "func"))
67 return DIFF_FUNCINFO;
68 return -1;
69 }
70
71 static int parse_dirstat_params(struct diff_options *options, const char *params,
72 struct strbuf *errmsg)
73 {
74 const char *p = params;
75 int p_len, ret = 0;
76
77 while (*p) {
78 p_len = strchrnul(p, ',') - p;
79 if (!memcmp(p, "changes", p_len)) {
80 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
81 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
82 } else if (!memcmp(p, "lines", p_len)) {
83 DIFF_OPT_SET(options, DIRSTAT_BY_LINE);
84 DIFF_OPT_CLR(options, DIRSTAT_BY_FILE);
85 } else if (!memcmp(p, "files", p_len)) {
86 DIFF_OPT_CLR(options, DIRSTAT_BY_LINE);
87 DIFF_OPT_SET(options, DIRSTAT_BY_FILE);
88 } else if (!memcmp(p, "noncumulative", p_len)) {
89 DIFF_OPT_CLR(options, DIRSTAT_CUMULATIVE);
90 } else if (!memcmp(p, "cumulative", p_len)) {
91 DIFF_OPT_SET(options, DIRSTAT_CUMULATIVE);
92 } else if (isdigit(*p)) {
93 char *end;
94 int permille = strtoul(p, &end, 10) * 10;
95 if (*end == '.' && isdigit(*++end)) {
96 /* only use first digit */
97 permille += *end - '0';
98 /* .. and ignore any further digits */
99 while (isdigit(*++end))
100 ; /* nothing */
101 }
102 if (end - p == p_len)
103 options->dirstat_permille = permille;
104 else {
105 strbuf_addf(errmsg, _(" Failed to parse dirstat cut-off percentage '%.*s'\n"),
106 p_len, p);
107 ret++;
108 }
109 } else {
110 strbuf_addf(errmsg, _(" Unknown dirstat parameter '%.*s'\n"),
111 p_len, p);
112 ret++;
113 }
114
115 p += p_len;
116
117 if (*p)
118 p++; /* more parameters, swallow separator */
119 }
120 return ret;
121 }
122
123 static int git_config_rename(const char *var, const char *value)
124 {
125 if (!value)
126 return DIFF_DETECT_RENAME;
127 if (!strcasecmp(value, "copies") || !strcasecmp(value, "copy"))
128 return DIFF_DETECT_COPY;
129 return git_config_bool(var,value) ? DIFF_DETECT_RENAME : 0;
130 }
131
132 /*
133 * These are to give UI layer defaults.
134 * The core-level commands such as git-diff-files should
135 * never be affected by the setting of diff.renames
136 * the user happens to have in the configuration file.
137 */
138 int git_diff_ui_config(const char *var, const char *value, void *cb)
139 {
140 if (!strcmp(var, "diff.color") || !strcmp(var, "color.diff")) {
141 diff_use_color_default = git_config_colorbool(var, value);
142 return 0;
143 }
144 if (!strcmp(var, "diff.renames")) {
145 diff_detect_rename_default = git_config_rename(var, value);
146 return 0;
147 }
148 if (!strcmp(var, "diff.autorefreshindex")) {
149 diff_auto_refresh_index = git_config_bool(var, value);
150 return 0;
151 }
152 if (!strcmp(var, "diff.mnemonicprefix")) {
153 diff_mnemonic_prefix = git_config_bool(var, value);
154 return 0;
155 }
156 if (!strcmp(var, "diff.noprefix")) {
157 diff_no_prefix = git_config_bool(var, value);
158 return 0;
159 }
160 if (!strcmp(var, "diff.statgraphwidth")) {
161 diff_stat_graph_width = git_config_int(var, value);
162 return 0;
163 }
164 if (!strcmp(var, "diff.external"))
165 return git_config_string(&external_diff_cmd_cfg, var, value);
166 if (!strcmp(var, "diff.wordregex"))
167 return git_config_string(&diff_word_regex_cfg, var, value);
168
169 if (!strcmp(var, "diff.ignoresubmodules"))
170 handle_ignore_submodules_arg(&default_diff_options, value);
171
172 if (git_color_config(var, value, cb) < 0)
173 return -1;
174
175 return git_diff_basic_config(var, value, cb);
176 }
177
178 int git_diff_basic_config(const char *var, const char *value, void *cb)
179 {
180 if (!strcmp(var, "diff.renamelimit")) {
181 diff_rename_limit_default = git_config_int(var, value);
182 return 0;
183 }
184
185 if (userdiff_config(var, value) < 0)
186 return -1;
187
188 if (!prefixcmp(var, "diff.color.") || !prefixcmp(var, "color.diff.")) {
189 int slot = parse_diff_color_slot(var, 11);
190 if (slot < 0)
191 return 0;
192 if (!value)
193 return config_error_nonbool(var);
194 color_parse(value, var, diff_colors[slot]);
195 return 0;
196 }
197
198 /* like GNU diff's --suppress-blank-empty option */
199 if (!strcmp(var, "diff.suppressblankempty") ||
200 /* for backwards compatibility */
201 !strcmp(var, "diff.suppress-blank-empty")) {
202 diff_suppress_blank_empty = git_config_bool(var, value);
203 return 0;
204 }
205
206 if (!strcmp(var, "diff.dirstat")) {
207 struct strbuf errmsg = STRBUF_INIT;
208 default_diff_options.dirstat_permille = diff_dirstat_permille_default;
209 if (parse_dirstat_params(&default_diff_options, value, &errmsg))
210 warning(_("Found errors in 'diff.dirstat' config variable:\n%s"),
211 errmsg.buf);
212 strbuf_release(&errmsg);
213 diff_dirstat_permille_default = default_diff_options.dirstat_permille;
214 return 0;
215 }
216
217 if (!prefixcmp(var, "submodule."))
218 return parse_submodule_config_option(var, value);
219
220 return git_default_config(var, value, cb);
221 }
222
223 static char *quote_two(const char *one, const char *two)
224 {
225 int need_one = quote_c_style(one, NULL, NULL, 1);
226 int need_two = quote_c_style(two, NULL, NULL, 1);
227 struct strbuf res = STRBUF_INIT;
228
229 if (need_one + need_two) {
230 strbuf_addch(&res, '"');
231 quote_c_style(one, &res, NULL, 1);
232 quote_c_style(two, &res, NULL, 1);
233 strbuf_addch(&res, '"');
234 } else {
235 strbuf_addstr(&res, one);
236 strbuf_addstr(&res, two);
237 }
238 return strbuf_detach(&res, NULL);
239 }
240
241 static const char *external_diff(void)
242 {
243 static const char *external_diff_cmd = NULL;
244 static int done_preparing = 0;
245
246 if (done_preparing)
247 return external_diff_cmd;
248 external_diff_cmd = getenv("GIT_EXTERNAL_DIFF");
249 if (!external_diff_cmd)
250 external_diff_cmd = external_diff_cmd_cfg;
251 done_preparing = 1;
252 return external_diff_cmd;
253 }
254
255 static struct diff_tempfile {
256 const char *name; /* filename external diff should read from */
257 char hex[41];
258 char mode[10];
259 char tmp_path[PATH_MAX];
260 } diff_temp[2];
261
262 typedef unsigned long (*sane_truncate_fn)(char *line, unsigned long len);
263
264 struct emit_callback {
265 int color_diff;
266 unsigned ws_rule;
267 int blank_at_eof_in_preimage;
268 int blank_at_eof_in_postimage;
269 int lno_in_preimage;
270 int lno_in_postimage;
271 sane_truncate_fn truncate;
272 const char **label_path;
273 struct diff_words_data *diff_words;
274 struct diff_options *opt;
275 int *found_changesp;
276 struct strbuf *header;
277 };
278
279 static int count_lines(const char *data, int size)
280 {
281 int count, ch, completely_empty = 1, nl_just_seen = 0;
282 count = 0;
283 while (0 < size--) {
284 ch = *data++;
285 if (ch == '\n') {
286 count++;
287 nl_just_seen = 1;
288 completely_empty = 0;
289 }
290 else {
291 nl_just_seen = 0;
292 completely_empty = 0;
293 }
294 }
295 if (completely_empty)
296 return 0;
297 if (!nl_just_seen)
298 count++; /* no trailing newline */
299 return count;
300 }
301
302 static int fill_mmfile(mmfile_t *mf, struct diff_filespec *one)
303 {
304 if (!DIFF_FILE_VALID(one)) {
305 mf->ptr = (char *)""; /* does not matter */
306 mf->size = 0;
307 return 0;
308 }
309 else if (diff_populate_filespec(one, 0))
310 return -1;
311
312 mf->ptr = one->data;
313 mf->size = one->size;
314 return 0;
315 }
316
317 /* like fill_mmfile, but only for size, so we can avoid retrieving blob */
318 static unsigned long diff_filespec_size(struct diff_filespec *one)
319 {
320 if (!DIFF_FILE_VALID(one))
321 return 0;
322 diff_populate_filespec(one, 1);
323 return one->size;
324 }
325
326 static int count_trailing_blank(mmfile_t *mf, unsigned ws_rule)
327 {
328 char *ptr = mf->ptr;
329 long size = mf->size;
330 int cnt = 0;
331
332 if (!size)
333 return cnt;
334 ptr += size - 1; /* pointing at the very end */
335 if (*ptr != '\n')
336 ; /* incomplete line */
337 else
338 ptr--; /* skip the last LF */
339 while (mf->ptr < ptr) {
340 char *prev_eol;
341 for (prev_eol = ptr; mf->ptr <= prev_eol; prev_eol--)
342 if (*prev_eol == '\n')
343 break;
344 if (!ws_blank_line(prev_eol + 1, ptr - prev_eol, ws_rule))
345 break;
346 cnt++;
347 ptr = prev_eol - 1;
348 }
349 return cnt;
350 }
351
352 static void check_blank_at_eof(mmfile_t *mf1, mmfile_t *mf2,
353 struct emit_callback *ecbdata)
354 {
355 int l1, l2, at;
356 unsigned ws_rule = ecbdata->ws_rule;
357 l1 = count_trailing_blank(mf1, ws_rule);
358 l2 = count_trailing_blank(mf2, ws_rule);
359 if (l2 <= l1) {
360 ecbdata->blank_at_eof_in_preimage = 0;
361 ecbdata->blank_at_eof_in_postimage = 0;
362 return;
363 }
364 at = count_lines(mf1->ptr, mf1->size);
365 ecbdata->blank_at_eof_in_preimage = (at - l1) + 1;
366
367 at = count_lines(mf2->ptr, mf2->size);
368 ecbdata->blank_at_eof_in_postimage = (at - l2) + 1;
369 }
370
371 static void emit_line_0(struct diff_options *o, const char *set, const char *reset,
372 int first, const char *line, int len)
373 {
374 int has_trailing_newline, has_trailing_carriage_return;
375 int nofirst;
376 FILE *file = o->file;
377
378 if (o->output_prefix) {
379 struct strbuf *msg = NULL;
380 msg = o->output_prefix(o, o->output_prefix_data);
381 assert(msg);
382 fwrite(msg->buf, msg->len, 1, file);
383 }
384
385 if (len == 0) {
386 has_trailing_newline = (first == '\n');
387 has_trailing_carriage_return = (!has_trailing_newline &&
388 (first == '\r'));
389 nofirst = has_trailing_newline || has_trailing_carriage_return;
390 } else {
391 has_trailing_newline = (len > 0 && line[len-1] == '\n');
392 if (has_trailing_newline)
393 len--;
394 has_trailing_carriage_return = (len > 0 && line[len-1] == '\r');
395 if (has_trailing_carriage_return)
396 len--;
397 nofirst = 0;
398 }
399
400 if (len || !nofirst) {
401 fputs(set, file);
402 if (!nofirst)
403 fputc(first, file);
404 fwrite(line, len, 1, file);
405 fputs(reset, file);
406 }
407 if (has_trailing_carriage_return)
408 fputc('\r', file);
409 if (has_trailing_newline)
410 fputc('\n', file);
411 }
412
413 static void emit_line(struct diff_options *o, const char *set, const char *reset,
414 const char *line, int len)
415 {
416 emit_line_0(o, set, reset, line[0], line+1, len-1);
417 }
418
419 static int new_blank_line_at_eof(struct emit_callback *ecbdata, const char *line, int len)
420 {
421 if (!((ecbdata->ws_rule & WS_BLANK_AT_EOF) &&
422 ecbdata->blank_at_eof_in_preimage &&
423 ecbdata->blank_at_eof_in_postimage &&
424 ecbdata->blank_at_eof_in_preimage <= ecbdata->lno_in_preimage &&
425 ecbdata->blank_at_eof_in_postimage <= ecbdata->lno_in_postimage))
426 return 0;
427 return ws_blank_line(line, len, ecbdata->ws_rule);
428 }
429
430 static void emit_add_line(const char *reset,
431 struct emit_callback *ecbdata,
432 const char *line, int len)
433 {
434 const char *ws = diff_get_color(ecbdata->color_diff, DIFF_WHITESPACE);
435 const char *set = diff_get_color(ecbdata->color_diff, DIFF_FILE_NEW);
436
437 if (!*ws)
438 emit_line_0(ecbdata->opt, set, reset, '+', line, len);
439 else if (new_blank_line_at_eof(ecbdata, line, len))
440 /* Blank line at EOF - paint '+' as well */
441 emit_line_0(ecbdata->opt, ws, reset, '+', line, len);
442 else {
443 /* Emit just the prefix, then the rest. */
444 emit_line_0(ecbdata->opt, set, reset, '+', "", 0);
445 ws_check_emit(line, len, ecbdata->ws_rule,
446 ecbdata->opt->file, set, reset, ws);
447 }
448 }
449
450 static void emit_hunk_header(struct emit_callback *ecbdata,
451 const char *line, int len)
452 {
453 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
454 const char *frag = diff_get_color(ecbdata->color_diff, DIFF_FRAGINFO);
455 const char *func = diff_get_color(ecbdata->color_diff, DIFF_FUNCINFO);
456 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
457 static const char atat[2] = { '@', '@' };
458 const char *cp, *ep;
459 struct strbuf msgbuf = STRBUF_INIT;
460 int org_len = len;
461 int i = 1;
462
463 /*
464 * As a hunk header must begin with "@@ -<old>, +<new> @@",
465 * it always is at least 10 bytes long.
466 */
467 if (len < 10 ||
468 memcmp(line, atat, 2) ||
469 !(ep = memmem(line + 2, len - 2, atat, 2))) {
470 emit_line(ecbdata->opt, plain, reset, line, len);
471 return;
472 }
473 ep += 2; /* skip over @@ */
474
475 /* The hunk header in fraginfo color */
476 strbuf_add(&msgbuf, frag, strlen(frag));
477 strbuf_add(&msgbuf, line, ep - line);
478 strbuf_add(&msgbuf, reset, strlen(reset));
479
480 /*
481 * trailing "\r\n"
482 */
483 for ( ; i < 3; i++)
484 if (line[len - i] == '\r' || line[len - i] == '\n')
485 len--;
486
487 /* blank before the func header */
488 for (cp = ep; ep - line < len; ep++)
489 if (*ep != ' ' && *ep != '\t')
490 break;
491 if (ep != cp) {
492 strbuf_add(&msgbuf, plain, strlen(plain));
493 strbuf_add(&msgbuf, cp, ep - cp);
494 strbuf_add(&msgbuf, reset, strlen(reset));
495 }
496
497 if (ep < line + len) {
498 strbuf_add(&msgbuf, func, strlen(func));
499 strbuf_add(&msgbuf, ep, line + len - ep);
500 strbuf_add(&msgbuf, reset, strlen(reset));
501 }
502
503 strbuf_add(&msgbuf, line + len, org_len - len);
504 emit_line(ecbdata->opt, "", "", msgbuf.buf, msgbuf.len);
505 strbuf_release(&msgbuf);
506 }
507
508 static struct diff_tempfile *claim_diff_tempfile(void) {
509 int i;
510 for (i = 0; i < ARRAY_SIZE(diff_temp); i++)
511 if (!diff_temp[i].name)
512 return diff_temp + i;
513 die("BUG: diff is failing to clean up its tempfiles");
514 }
515
516 static int remove_tempfile_installed;
517
518 static void remove_tempfile(void)
519 {
520 int i;
521 for (i = 0; i < ARRAY_SIZE(diff_temp); i++) {
522 if (diff_temp[i].name == diff_temp[i].tmp_path)
523 unlink_or_warn(diff_temp[i].name);
524 diff_temp[i].name = NULL;
525 }
526 }
527
528 static void remove_tempfile_on_signal(int signo)
529 {
530 remove_tempfile();
531 sigchain_pop(signo);
532 raise(signo);
533 }
534
535 static void print_line_count(FILE *file, int count)
536 {
537 switch (count) {
538 case 0:
539 fprintf(file, "0,0");
540 break;
541 case 1:
542 fprintf(file, "1");
543 break;
544 default:
545 fprintf(file, "1,%d", count);
546 break;
547 }
548 }
549
550 static void emit_rewrite_lines(struct emit_callback *ecb,
551 int prefix, const char *data, int size)
552 {
553 const char *endp = NULL;
554 static const char *nneof = " No newline at end of file\n";
555 const char *old = diff_get_color(ecb->color_diff, DIFF_FILE_OLD);
556 const char *reset = diff_get_color(ecb->color_diff, DIFF_RESET);
557
558 while (0 < size) {
559 int len;
560
561 endp = memchr(data, '\n', size);
562 len = endp ? (endp - data + 1) : size;
563 if (prefix != '+') {
564 ecb->lno_in_preimage++;
565 emit_line_0(ecb->opt, old, reset, '-',
566 data, len);
567 } else {
568 ecb->lno_in_postimage++;
569 emit_add_line(reset, ecb, data, len);
570 }
571 size -= len;
572 data += len;
573 }
574 if (!endp) {
575 const char *plain = diff_get_color(ecb->color_diff,
576 DIFF_PLAIN);
577 emit_line_0(ecb->opt, plain, reset, '\\',
578 nneof, strlen(nneof));
579 }
580 }
581
582 static void emit_rewrite_diff(const char *name_a,
583 const char *name_b,
584 struct diff_filespec *one,
585 struct diff_filespec *two,
586 struct userdiff_driver *textconv_one,
587 struct userdiff_driver *textconv_two,
588 struct diff_options *o)
589 {
590 int lc_a, lc_b;
591 const char *name_a_tab, *name_b_tab;
592 const char *metainfo = diff_get_color(o->use_color, DIFF_METAINFO);
593 const char *fraginfo = diff_get_color(o->use_color, DIFF_FRAGINFO);
594 const char *reset = diff_get_color(o->use_color, DIFF_RESET);
595 static struct strbuf a_name = STRBUF_INIT, b_name = STRBUF_INIT;
596 const char *a_prefix, *b_prefix;
597 char *data_one, *data_two;
598 size_t size_one, size_two;
599 struct emit_callback ecbdata;
600 char *line_prefix = "";
601 struct strbuf *msgbuf;
602
603 if (o && o->output_prefix) {
604 msgbuf = o->output_prefix(o, o->output_prefix_data);
605 line_prefix = msgbuf->buf;
606 }
607
608 if (diff_mnemonic_prefix && DIFF_OPT_TST(o, REVERSE_DIFF)) {
609 a_prefix = o->b_prefix;
610 b_prefix = o->a_prefix;
611 } else {
612 a_prefix = o->a_prefix;
613 b_prefix = o->b_prefix;
614 }
615
616 name_a += (*name_a == '/');
617 name_b += (*name_b == '/');
618 name_a_tab = strchr(name_a, ' ') ? "\t" : "";
619 name_b_tab = strchr(name_b, ' ') ? "\t" : "";
620
621 strbuf_reset(&a_name);
622 strbuf_reset(&b_name);
623 quote_two_c_style(&a_name, a_prefix, name_a, 0);
624 quote_two_c_style(&b_name, b_prefix, name_b, 0);
625
626 size_one = fill_textconv(textconv_one, one, &data_one);
627 size_two = fill_textconv(textconv_two, two, &data_two);
628
629 memset(&ecbdata, 0, sizeof(ecbdata));
630 ecbdata.color_diff = want_color(o->use_color);
631 ecbdata.found_changesp = &o->found_changes;
632 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
633 ecbdata.opt = o;
634 if (ecbdata.ws_rule & WS_BLANK_AT_EOF) {
635 mmfile_t mf1, mf2;
636 mf1.ptr = (char *)data_one;
637 mf2.ptr = (char *)data_two;
638 mf1.size = size_one;
639 mf2.size = size_two;
640 check_blank_at_eof(&mf1, &mf2, &ecbdata);
641 }
642 ecbdata.lno_in_preimage = 1;
643 ecbdata.lno_in_postimage = 1;
644
645 lc_a = count_lines(data_one, size_one);
646 lc_b = count_lines(data_two, size_two);
647 fprintf(o->file,
648 "%s%s--- %s%s%s\n%s%s+++ %s%s%s\n%s%s@@ -",
649 line_prefix, metainfo, a_name.buf, name_a_tab, reset,
650 line_prefix, metainfo, b_name.buf, name_b_tab, reset,
651 line_prefix, fraginfo);
652 if (!o->irreversible_delete)
653 print_line_count(o->file, lc_a);
654 else
655 fprintf(o->file, "?,?");
656 fprintf(o->file, " +");
657 print_line_count(o->file, lc_b);
658 fprintf(o->file, " @@%s\n", reset);
659 if (lc_a && !o->irreversible_delete)
660 emit_rewrite_lines(&ecbdata, '-', data_one, size_one);
661 if (lc_b)
662 emit_rewrite_lines(&ecbdata, '+', data_two, size_two);
663 if (textconv_one)
664 free((char *)data_one);
665 if (textconv_two)
666 free((char *)data_two);
667 }
668
669 struct diff_words_buffer {
670 mmfile_t text;
671 long alloc;
672 struct diff_words_orig {
673 const char *begin, *end;
674 } *orig;
675 int orig_nr, orig_alloc;
676 };
677
678 static void diff_words_append(char *line, unsigned long len,
679 struct diff_words_buffer *buffer)
680 {
681 ALLOC_GROW(buffer->text.ptr, buffer->text.size + len, buffer->alloc);
682 line++;
683 len--;
684 memcpy(buffer->text.ptr + buffer->text.size, line, len);
685 buffer->text.size += len;
686 buffer->text.ptr[buffer->text.size] = '\0';
687 }
688
689 struct diff_words_style_elem {
690 const char *prefix;
691 const char *suffix;
692 const char *color; /* NULL; filled in by the setup code if
693 * color is enabled */
694 };
695
696 struct diff_words_style {
697 enum diff_words_type type;
698 struct diff_words_style_elem new, old, ctx;
699 const char *newline;
700 };
701
702 static struct diff_words_style diff_words_styles[] = {
703 { DIFF_WORDS_PORCELAIN, {"+", "\n"}, {"-", "\n"}, {" ", "\n"}, "~\n" },
704 { DIFF_WORDS_PLAIN, {"{+", "+}"}, {"[-", "-]"}, {"", ""}, "\n" },
705 { DIFF_WORDS_COLOR, {"", ""}, {"", ""}, {"", ""}, "\n" }
706 };
707
708 struct diff_words_data {
709 struct diff_words_buffer minus, plus;
710 const char *current_plus;
711 int last_minus;
712 struct diff_options *opt;
713 regex_t *word_regex;
714 enum diff_words_type type;
715 struct diff_words_style *style;
716 };
717
718 static int fn_out_diff_words_write_helper(FILE *fp,
719 struct diff_words_style_elem *st_el,
720 const char *newline,
721 size_t count, const char *buf,
722 const char *line_prefix)
723 {
724 int print = 0;
725
726 while (count) {
727 char *p = memchr(buf, '\n', count);
728 if (print)
729 fputs(line_prefix, fp);
730 if (p != buf) {
731 if (st_el->color && fputs(st_el->color, fp) < 0)
732 return -1;
733 if (fputs(st_el->prefix, fp) < 0 ||
734 fwrite(buf, p ? p - buf : count, 1, fp) != 1 ||
735 fputs(st_el->suffix, fp) < 0)
736 return -1;
737 if (st_el->color && *st_el->color
738 && fputs(GIT_COLOR_RESET, fp) < 0)
739 return -1;
740 }
741 if (!p)
742 return 0;
743 if (fputs(newline, fp) < 0)
744 return -1;
745 count -= p + 1 - buf;
746 buf = p + 1;
747 print = 1;
748 }
749 return 0;
750 }
751
752 /*
753 * '--color-words' algorithm can be described as:
754 *
755 * 1. collect a the minus/plus lines of a diff hunk, divided into
756 * minus-lines and plus-lines;
757 *
758 * 2. break both minus-lines and plus-lines into words and
759 * place them into two mmfile_t with one word for each line;
760 *
761 * 3. use xdiff to run diff on the two mmfile_t to get the words level diff;
762 *
763 * And for the common parts of the both file, we output the plus side text.
764 * diff_words->current_plus is used to trace the current position of the plus file
765 * which printed. diff_words->last_minus is used to trace the last minus word
766 * printed.
767 *
768 * For '--graph' to work with '--color-words', we need to output the graph prefix
769 * on each line of color words output. Generally, there are two conditions on
770 * which we should output the prefix.
771 *
772 * 1. diff_words->last_minus == 0 &&
773 * diff_words->current_plus == diff_words->plus.text.ptr
774 *
775 * that is: the plus text must start as a new line, and if there is no minus
776 * word printed, a graph prefix must be printed.
777 *
778 * 2. diff_words->current_plus > diff_words->plus.text.ptr &&
779 * *(diff_words->current_plus - 1) == '\n'
780 *
781 * that is: a graph prefix must be printed following a '\n'
782 */
783 static int color_words_output_graph_prefix(struct diff_words_data *diff_words)
784 {
785 if ((diff_words->last_minus == 0 &&
786 diff_words->current_plus == diff_words->plus.text.ptr) ||
787 (diff_words->current_plus > diff_words->plus.text.ptr &&
788 *(diff_words->current_plus - 1) == '\n')) {
789 return 1;
790 } else {
791 return 0;
792 }
793 }
794
795 static void fn_out_diff_words_aux(void *priv, char *line, unsigned long len)
796 {
797 struct diff_words_data *diff_words = priv;
798 struct diff_words_style *style = diff_words->style;
799 int minus_first, minus_len, plus_first, plus_len;
800 const char *minus_begin, *minus_end, *plus_begin, *plus_end;
801 struct diff_options *opt = diff_words->opt;
802 struct strbuf *msgbuf;
803 char *line_prefix = "";
804
805 if (line[0] != '@' || parse_hunk_header(line, len,
806 &minus_first, &minus_len, &plus_first, &plus_len))
807 return;
808
809 assert(opt);
810 if (opt->output_prefix) {
811 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
812 line_prefix = msgbuf->buf;
813 }
814
815 /* POSIX requires that first be decremented by one if len == 0... */
816 if (minus_len) {
817 minus_begin = diff_words->minus.orig[minus_first].begin;
818 minus_end =
819 diff_words->minus.orig[minus_first + minus_len - 1].end;
820 } else
821 minus_begin = minus_end =
822 diff_words->minus.orig[minus_first].end;
823
824 if (plus_len) {
825 plus_begin = diff_words->plus.orig[plus_first].begin;
826 plus_end = diff_words->plus.orig[plus_first + plus_len - 1].end;
827 } else
828 plus_begin = plus_end = diff_words->plus.orig[plus_first].end;
829
830 if (color_words_output_graph_prefix(diff_words)) {
831 fputs(line_prefix, diff_words->opt->file);
832 }
833 if (diff_words->current_plus != plus_begin) {
834 fn_out_diff_words_write_helper(diff_words->opt->file,
835 &style->ctx, style->newline,
836 plus_begin - diff_words->current_plus,
837 diff_words->current_plus, line_prefix);
838 if (*(plus_begin - 1) == '\n')
839 fputs(line_prefix, diff_words->opt->file);
840 }
841 if (minus_begin != minus_end) {
842 fn_out_diff_words_write_helper(diff_words->opt->file,
843 &style->old, style->newline,
844 minus_end - minus_begin, minus_begin,
845 line_prefix);
846 }
847 if (plus_begin != plus_end) {
848 fn_out_diff_words_write_helper(diff_words->opt->file,
849 &style->new, style->newline,
850 plus_end - plus_begin, plus_begin,
851 line_prefix);
852 }
853
854 diff_words->current_plus = plus_end;
855 diff_words->last_minus = minus_first;
856 }
857
858 /* This function starts looking at *begin, and returns 0 iff a word was found. */
859 static int find_word_boundaries(mmfile_t *buffer, regex_t *word_regex,
860 int *begin, int *end)
861 {
862 if (word_regex && *begin < buffer->size) {
863 regmatch_t match[1];
864 if (!regexec(word_regex, buffer->ptr + *begin, 1, match, 0)) {
865 char *p = memchr(buffer->ptr + *begin + match[0].rm_so,
866 '\n', match[0].rm_eo - match[0].rm_so);
867 *end = p ? p - buffer->ptr : match[0].rm_eo + *begin;
868 *begin += match[0].rm_so;
869 return *begin >= *end;
870 }
871 return -1;
872 }
873
874 /* find the next word */
875 while (*begin < buffer->size && isspace(buffer->ptr[*begin]))
876 (*begin)++;
877 if (*begin >= buffer->size)
878 return -1;
879
880 /* find the end of the word */
881 *end = *begin + 1;
882 while (*end < buffer->size && !isspace(buffer->ptr[*end]))
883 (*end)++;
884
885 return 0;
886 }
887
888 /*
889 * This function splits the words in buffer->text, stores the list with
890 * newline separator into out, and saves the offsets of the original words
891 * in buffer->orig.
892 */
893 static void diff_words_fill(struct diff_words_buffer *buffer, mmfile_t *out,
894 regex_t *word_regex)
895 {
896 int i, j;
897 long alloc = 0;
898
899 out->size = 0;
900 out->ptr = NULL;
901
902 /* fake an empty "0th" word */
903 ALLOC_GROW(buffer->orig, 1, buffer->orig_alloc);
904 buffer->orig[0].begin = buffer->orig[0].end = buffer->text.ptr;
905 buffer->orig_nr = 1;
906
907 for (i = 0; i < buffer->text.size; i++) {
908 if (find_word_boundaries(&buffer->text, word_regex, &i, &j))
909 return;
910
911 /* store original boundaries */
912 ALLOC_GROW(buffer->orig, buffer->orig_nr + 1,
913 buffer->orig_alloc);
914 buffer->orig[buffer->orig_nr].begin = buffer->text.ptr + i;
915 buffer->orig[buffer->orig_nr].end = buffer->text.ptr + j;
916 buffer->orig_nr++;
917
918 /* store one word */
919 ALLOC_GROW(out->ptr, out->size + j - i + 1, alloc);
920 memcpy(out->ptr + out->size, buffer->text.ptr + i, j - i);
921 out->ptr[out->size + j - i] = '\n';
922 out->size += j - i + 1;
923
924 i = j - 1;
925 }
926 }
927
928 /* this executes the word diff on the accumulated buffers */
929 static void diff_words_show(struct diff_words_data *diff_words)
930 {
931 xpparam_t xpp;
932 xdemitconf_t xecfg;
933 mmfile_t minus, plus;
934 struct diff_words_style *style = diff_words->style;
935
936 struct diff_options *opt = diff_words->opt;
937 struct strbuf *msgbuf;
938 char *line_prefix = "";
939
940 assert(opt);
941 if (opt->output_prefix) {
942 msgbuf = opt->output_prefix(opt, opt->output_prefix_data);
943 line_prefix = msgbuf->buf;
944 }
945
946 /* special case: only removal */
947 if (!diff_words->plus.text.size) {
948 fputs(line_prefix, diff_words->opt->file);
949 fn_out_diff_words_write_helper(diff_words->opt->file,
950 &style->old, style->newline,
951 diff_words->minus.text.size,
952 diff_words->minus.text.ptr, line_prefix);
953 diff_words->minus.text.size = 0;
954 return;
955 }
956
957 diff_words->current_plus = diff_words->plus.text.ptr;
958 diff_words->last_minus = 0;
959
960 memset(&xpp, 0, sizeof(xpp));
961 memset(&xecfg, 0, sizeof(xecfg));
962 diff_words_fill(&diff_words->minus, &minus, diff_words->word_regex);
963 diff_words_fill(&diff_words->plus, &plus, diff_words->word_regex);
964 xpp.flags = 0;
965 /* as only the hunk header will be parsed, we need a 0-context */
966 xecfg.ctxlen = 0;
967 xdi_diff_outf(&minus, &plus, fn_out_diff_words_aux, diff_words,
968 &xpp, &xecfg);
969 free(minus.ptr);
970 free(plus.ptr);
971 if (diff_words->current_plus != diff_words->plus.text.ptr +
972 diff_words->plus.text.size) {
973 if (color_words_output_graph_prefix(diff_words))
974 fputs(line_prefix, diff_words->opt->file);
975 fn_out_diff_words_write_helper(diff_words->opt->file,
976 &style->ctx, style->newline,
977 diff_words->plus.text.ptr + diff_words->plus.text.size
978 - diff_words->current_plus, diff_words->current_plus,
979 line_prefix);
980 }
981 diff_words->minus.text.size = diff_words->plus.text.size = 0;
982 }
983
984 /* In "color-words" mode, show word-diff of words accumulated in the buffer */
985 static void diff_words_flush(struct emit_callback *ecbdata)
986 {
987 if (ecbdata->diff_words->minus.text.size ||
988 ecbdata->diff_words->plus.text.size)
989 diff_words_show(ecbdata->diff_words);
990 }
991
992 static void diff_filespec_load_driver(struct diff_filespec *one)
993 {
994 /* Use already-loaded driver */
995 if (one->driver)
996 return;
997
998 if (S_ISREG(one->mode))
999 one->driver = userdiff_find_by_path(one->path);
1000
1001 /* Fallback to default settings */
1002 if (!one->driver)
1003 one->driver = userdiff_find_by_name("default");
1004 }
1005
1006 static const char *userdiff_word_regex(struct diff_filespec *one)
1007 {
1008 diff_filespec_load_driver(one);
1009 return one->driver->word_regex;
1010 }
1011
1012 static void init_diff_words_data(struct emit_callback *ecbdata,
1013 struct diff_options *orig_opts,
1014 struct diff_filespec *one,
1015 struct diff_filespec *two)
1016 {
1017 int i;
1018 struct diff_options *o = xmalloc(sizeof(struct diff_options));
1019 memcpy(o, orig_opts, sizeof(struct diff_options));
1020
1021 ecbdata->diff_words =
1022 xcalloc(1, sizeof(struct diff_words_data));
1023 ecbdata->diff_words->type = o->word_diff;
1024 ecbdata->diff_words->opt = o;
1025 if (!o->word_regex)
1026 o->word_regex = userdiff_word_regex(one);
1027 if (!o->word_regex)
1028 o->word_regex = userdiff_word_regex(two);
1029 if (!o->word_regex)
1030 o->word_regex = diff_word_regex_cfg;
1031 if (o->word_regex) {
1032 ecbdata->diff_words->word_regex = (regex_t *)
1033 xmalloc(sizeof(regex_t));
1034 if (regcomp(ecbdata->diff_words->word_regex,
1035 o->word_regex,
1036 REG_EXTENDED | REG_NEWLINE))
1037 die ("Invalid regular expression: %s",
1038 o->word_regex);
1039 }
1040 for (i = 0; i < ARRAY_SIZE(diff_words_styles); i++) {
1041 if (o->word_diff == diff_words_styles[i].type) {
1042 ecbdata->diff_words->style =
1043 &diff_words_styles[i];
1044 break;
1045 }
1046 }
1047 if (want_color(o->use_color)) {
1048 struct diff_words_style *st = ecbdata->diff_words->style;
1049 st->old.color = diff_get_color_opt(o, DIFF_FILE_OLD);
1050 st->new.color = diff_get_color_opt(o, DIFF_FILE_NEW);
1051 st->ctx.color = diff_get_color_opt(o, DIFF_PLAIN);
1052 }
1053 }
1054
1055 static void free_diff_words_data(struct emit_callback *ecbdata)
1056 {
1057 if (ecbdata->diff_words) {
1058 diff_words_flush(ecbdata);
1059 free (ecbdata->diff_words->opt);
1060 free (ecbdata->diff_words->minus.text.ptr);
1061 free (ecbdata->diff_words->minus.orig);
1062 free (ecbdata->diff_words->plus.text.ptr);
1063 free (ecbdata->diff_words->plus.orig);
1064 if (ecbdata->diff_words->word_regex) {
1065 regfree(ecbdata->diff_words->word_regex);
1066 free(ecbdata->diff_words->word_regex);
1067 }
1068 free(ecbdata->diff_words);
1069 ecbdata->diff_words = NULL;
1070 }
1071 }
1072
1073 const char *diff_get_color(int diff_use_color, enum color_diff ix)
1074 {
1075 if (want_color(diff_use_color))
1076 return diff_colors[ix];
1077 return "";
1078 }
1079
1080 static unsigned long sane_truncate_line(struct emit_callback *ecb, char *line, unsigned long len)
1081 {
1082 const char *cp;
1083 unsigned long allot;
1084 size_t l = len;
1085
1086 if (ecb->truncate)
1087 return ecb->truncate(line, len);
1088 cp = line;
1089 allot = l;
1090 while (0 < l) {
1091 (void) utf8_width(&cp, &l);
1092 if (!cp)
1093 break; /* truncated in the middle? */
1094 }
1095 return allot - l;
1096 }
1097
1098 static void find_lno(const char *line, struct emit_callback *ecbdata)
1099 {
1100 const char *p;
1101 ecbdata->lno_in_preimage = 0;
1102 ecbdata->lno_in_postimage = 0;
1103 p = strchr(line, '-');
1104 if (!p)
1105 return; /* cannot happen */
1106 ecbdata->lno_in_preimage = strtol(p + 1, NULL, 10);
1107 p = strchr(p, '+');
1108 if (!p)
1109 return; /* cannot happen */
1110 ecbdata->lno_in_postimage = strtol(p + 1, NULL, 10);
1111 }
1112
1113 static void fn_out_consume(void *priv, char *line, unsigned long len)
1114 {
1115 struct emit_callback *ecbdata = priv;
1116 const char *meta = diff_get_color(ecbdata->color_diff, DIFF_METAINFO);
1117 const char *plain = diff_get_color(ecbdata->color_diff, DIFF_PLAIN);
1118 const char *reset = diff_get_color(ecbdata->color_diff, DIFF_RESET);
1119 struct diff_options *o = ecbdata->opt;
1120 char *line_prefix = "";
1121 struct strbuf *msgbuf;
1122
1123 if (o && o->output_prefix) {
1124 msgbuf = o->output_prefix(o, o->output_prefix_data);
1125 line_prefix = msgbuf->buf;
1126 }
1127
1128 if (ecbdata->header) {
1129 fprintf(ecbdata->opt->file, "%s", ecbdata->header->buf);
1130 strbuf_reset(ecbdata->header);
1131 ecbdata->header = NULL;
1132 }
1133 *(ecbdata->found_changesp) = 1;
1134
1135 if (ecbdata->label_path[0]) {
1136 const char *name_a_tab, *name_b_tab;
1137
1138 name_a_tab = strchr(ecbdata->label_path[0], ' ') ? "\t" : "";
1139 name_b_tab = strchr(ecbdata->label_path[1], ' ') ? "\t" : "";
1140
1141 fprintf(ecbdata->opt->file, "%s%s--- %s%s%s\n",
1142 line_prefix, meta, ecbdata->label_path[0], reset, name_a_tab);
1143 fprintf(ecbdata->opt->file, "%s%s+++ %s%s%s\n",
1144 line_prefix, meta, ecbdata->label_path[1], reset, name_b_tab);
1145 ecbdata->label_path[0] = ecbdata->label_path[1] = NULL;
1146 }
1147
1148 if (diff_suppress_blank_empty
1149 && len == 2 && line[0] == ' ' && line[1] == '\n') {
1150 line[0] = '\n';
1151 len = 1;
1152 }
1153
1154 if (line[0] == '@') {
1155 if (ecbdata->diff_words)
1156 diff_words_flush(ecbdata);
1157 len = sane_truncate_line(ecbdata, line, len);
1158 find_lno(line, ecbdata);
1159 emit_hunk_header(ecbdata, line, len);
1160 if (line[len-1] != '\n')
1161 putc('\n', ecbdata->opt->file);
1162 return;
1163 }
1164
1165 if (len < 1) {
1166 emit_line(ecbdata->opt, reset, reset, line, len);
1167 if (ecbdata->diff_words
1168 && ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN)
1169 fputs("~\n", ecbdata->opt->file);
1170 return;
1171 }
1172
1173 if (ecbdata->diff_words) {
1174 if (line[0] == '-') {
1175 diff_words_append(line, len,
1176 &ecbdata->diff_words->minus);
1177 return;
1178 } else if (line[0] == '+') {
1179 diff_words_append(line, len,
1180 &ecbdata->diff_words->plus);
1181 return;
1182 } else if (!prefixcmp(line, "\\ ")) {
1183 /*
1184 * Eat the "no newline at eof" marker as if we
1185 * saw a "+" or "-" line with nothing on it,
1186 * and return without diff_words_flush() to
1187 * defer processing. If this is the end of
1188 * preimage, more "+" lines may come after it.
1189 */
1190 return;
1191 }
1192 diff_words_flush(ecbdata);
1193 if (ecbdata->diff_words->type == DIFF_WORDS_PORCELAIN) {
1194 emit_line(ecbdata->opt, plain, reset, line, len);
1195 fputs("~\n", ecbdata->opt->file);
1196 } else {
1197 /*
1198 * Skip the prefix character, if any. With
1199 * diff_suppress_blank_empty, there may be
1200 * none.
1201 */
1202 if (line[0] != '\n') {
1203 line++;
1204 len--;
1205 }
1206 emit_line(ecbdata->opt, plain, reset, line, len);
1207 }
1208 return;
1209 }
1210
1211 if (line[0] != '+') {
1212 const char *color =
1213 diff_get_color(ecbdata->color_diff,
1214 line[0] == '-' ? DIFF_FILE_OLD : DIFF_PLAIN);
1215 ecbdata->lno_in_preimage++;
1216 if (line[0] == ' ')
1217 ecbdata->lno_in_postimage++;
1218 emit_line(ecbdata->opt, color, reset, line, len);
1219 } else {
1220 ecbdata->lno_in_postimage++;
1221 emit_add_line(reset, ecbdata, line + 1, len - 1);
1222 }
1223 }
1224
1225 static char *pprint_rename(const char *a, const char *b)
1226 {
1227 const char *old = a;
1228 const char *new = b;
1229 struct strbuf name = STRBUF_INIT;
1230 int pfx_length, sfx_length;
1231 int len_a = strlen(a);
1232 int len_b = strlen(b);
1233 int a_midlen, b_midlen;
1234 int qlen_a = quote_c_style(a, NULL, NULL, 0);
1235 int qlen_b = quote_c_style(b, NULL, NULL, 0);
1236
1237 if (qlen_a || qlen_b) {
1238 quote_c_style(a, &name, NULL, 0);
1239 strbuf_addstr(&name, " => ");
1240 quote_c_style(b, &name, NULL, 0);
1241 return strbuf_detach(&name, NULL);
1242 }
1243
1244 /* Find common prefix */
1245 pfx_length = 0;
1246 while (*old && *new && *old == *new) {
1247 if (*old == '/')
1248 pfx_length = old - a + 1;
1249 old++;
1250 new++;
1251 }
1252
1253 /* Find common suffix */
1254 old = a + len_a;
1255 new = b + len_b;
1256 sfx_length = 0;
1257 while (a <= old && b <= new && *old == *new) {
1258 if (*old == '/')
1259 sfx_length = len_a - (old - a);
1260 old--;
1261 new--;
1262 }
1263
1264 /*
1265 * pfx{mid-a => mid-b}sfx
1266 * {pfx-a => pfx-b}sfx
1267 * pfx{sfx-a => sfx-b}
1268 * name-a => name-b
1269 */
1270 a_midlen = len_a - pfx_length - sfx_length;
1271 b_midlen = len_b - pfx_length - sfx_length;
1272 if (a_midlen < 0)
1273 a_midlen = 0;
1274 if (b_midlen < 0)
1275 b_midlen = 0;
1276
1277 strbuf_grow(&name, pfx_length + a_midlen + b_midlen + sfx_length + 7);
1278 if (pfx_length + sfx_length) {
1279 strbuf_add(&name, a, pfx_length);
1280 strbuf_addch(&name, '{');
1281 }
1282 strbuf_add(&name, a + pfx_length, a_midlen);
1283 strbuf_addstr(&name, " => ");
1284 strbuf_add(&name, b + pfx_length, b_midlen);
1285 if (pfx_length + sfx_length) {
1286 strbuf_addch(&name, '}');
1287 strbuf_add(&name, a + len_a - sfx_length, sfx_length);
1288 }
1289 return strbuf_detach(&name, NULL);
1290 }
1291
1292 struct diffstat_t {
1293 int nr;
1294 int alloc;
1295 struct diffstat_file {
1296 char *from_name;
1297 char *name;
1298 char *print_name;
1299 unsigned is_unmerged:1;
1300 unsigned is_binary:1;
1301 unsigned is_renamed:1;
1302 uintmax_t added, deleted;
1303 } **files;
1304 };
1305
1306 static struct diffstat_file *diffstat_add(struct diffstat_t *diffstat,
1307 const char *name_a,
1308 const char *name_b)
1309 {
1310 struct diffstat_file *x;
1311 x = xcalloc(sizeof (*x), 1);
1312 if (diffstat->nr == diffstat->alloc) {
1313 diffstat->alloc = alloc_nr(diffstat->alloc);
1314 diffstat->files = xrealloc(diffstat->files,
1315 diffstat->alloc * sizeof(x));
1316 }
1317 diffstat->files[diffstat->nr++] = x;
1318 if (name_b) {
1319 x->from_name = xstrdup(name_a);
1320 x->name = xstrdup(name_b);
1321 x->is_renamed = 1;
1322 }
1323 else {
1324 x->from_name = NULL;
1325 x->name = xstrdup(name_a);
1326 }
1327 return x;
1328 }
1329
1330 static void diffstat_consume(void *priv, char *line, unsigned long len)
1331 {
1332 struct diffstat_t *diffstat = priv;
1333 struct diffstat_file *x = diffstat->files[diffstat->nr - 1];
1334
1335 if (line[0] == '+')
1336 x->added++;
1337 else if (line[0] == '-')
1338 x->deleted++;
1339 }
1340
1341 const char mime_boundary_leader[] = "------------";
1342
1343 static int scale_linear(int it, int width, int max_change)
1344 {
1345 if (!it)
1346 return 0;
1347 /*
1348 * make sure that at least one '-' or '+' is printed if
1349 * there is any change to this path. The easiest way is to
1350 * scale linearly as if the alloted width is one column shorter
1351 * than it is, and then add 1 to the result.
1352 */
1353 return 1 + (it * (width - 1) / max_change);
1354 }
1355
1356 static void show_name(FILE *file,
1357 const char *prefix, const char *name, int len)
1358 {
1359 fprintf(file, " %s%-*s |", prefix, len, name);
1360 }
1361
1362 static void show_graph(FILE *file, char ch, int cnt, const char *set, const char *reset)
1363 {
1364 if (cnt <= 0)
1365 return;
1366 fprintf(file, "%s", set);
1367 while (cnt--)
1368 putc(ch, file);
1369 fprintf(file, "%s", reset);
1370 }
1371
1372 static void fill_print_name(struct diffstat_file *file)
1373 {
1374 char *pname;
1375
1376 if (file->print_name)
1377 return;
1378
1379 if (!file->is_renamed) {
1380 struct strbuf buf = STRBUF_INIT;
1381 if (quote_c_style(file->name, &buf, NULL, 0)) {
1382 pname = strbuf_detach(&buf, NULL);
1383 } else {
1384 pname = file->name;
1385 strbuf_release(&buf);
1386 }
1387 } else {
1388 pname = pprint_rename(file->from_name, file->name);
1389 }
1390 file->print_name = pname;
1391 }
1392
1393 int print_stat_summary(FILE *fp, int files, int insertions, int deletions)
1394 {
1395 struct strbuf sb = STRBUF_INIT;
1396 int ret;
1397
1398 if (!files) {
1399 assert(insertions == 0 && deletions == 0);
1400 return fputs(_(" 0 files changed\n"), fp);
1401 }
1402
1403 strbuf_addf(&sb,
1404 Q_(" %d file changed", " %d files changed", files),
1405 files);
1406
1407 /*
1408 * For binary diff, the caller may want to print "x files
1409 * changed" with insertions == 0 && deletions == 0.
1410 *
1411 * Not omitting "0 insertions(+), 0 deletions(-)" in this case
1412 * is probably less confusing (i.e skip over "2 files changed
1413 * but nothing about added/removed lines? Is this a bug in Git?").
1414 */
1415 if (insertions || deletions == 0) {
1416 /*
1417 * TRANSLATORS: "+" in (+) is a line addition marker;
1418 * do not translate it.
1419 */
1420 strbuf_addf(&sb,
1421 Q_(", %d insertion(+)", ", %d insertions(+)",
1422 insertions),
1423 insertions);
1424 }
1425
1426 if (deletions || insertions == 0) {
1427 /*
1428 * TRANSLATORS: "-" in (-) is a line removal marker;
1429 * do not translate it.
1430 */
1431 strbuf_addf(&sb,
1432 Q_(", %d deletion(-)", ", %d deletions(-)",
1433 deletions),
1434 deletions);
1435 }
1436 strbuf_addch(&sb, '\n');
1437 ret = fputs(sb.buf, fp);
1438 strbuf_release(&sb);
1439 return ret;
1440 }
1441
1442 static void show_stats(struct diffstat_t *data, struct diff_options *options)
1443 {
1444 int i, len, add, del, adds = 0, dels = 0;
1445 uintmax_t max_change = 0, max_len = 0;
1446 int total_files = data->nr, count;
1447 int width, name_width, graph_width, number_width = 0, bin_width = 0;
1448 const char *reset, *add_c, *del_c;
1449 const char *line_prefix = "";
1450 int extra_shown = 0;
1451 struct strbuf *msg = NULL;
1452
1453 if (data->nr == 0)
1454 return;
1455
1456 if (options->output_prefix) {
1457 msg = options->output_prefix(options, options->output_prefix_data);
1458 line_prefix = msg->buf;
1459 }
1460
1461 count = options->stat_count ? options->stat_count : data->nr;
1462
1463 reset = diff_get_color_opt(options, DIFF_RESET);
1464 add_c = diff_get_color_opt(options, DIFF_FILE_NEW);
1465 del_c = diff_get_color_opt(options, DIFF_FILE_OLD);
1466
1467 /*
1468 * Find the longest filename and max number of changes
1469 */
1470 for (i = 0; (i < count) && (i < data->nr); i++) {
1471 struct diffstat_file *file = data->files[i];
1472 uintmax_t change = file->added + file->deleted;
1473 if (!data->files[i]->is_renamed &&
1474 (change == 0)) {
1475 count++; /* not shown == room for one more */
1476 continue;
1477 }
1478 fill_print_name(file);
1479 len = strlen(file->print_name);
1480 if (max_len < len)
1481 max_len = len;
1482
1483 if (file->is_unmerged) {
1484 /* "Unmerged" is 8 characters */
1485 bin_width = bin_width < 8 ? 8 : bin_width;
1486 continue;
1487 }
1488 if (file->is_binary) {
1489 /* "Bin XXX -> YYY bytes" */
1490 int w = 14 + decimal_width(file->added)
1491 + decimal_width(file->deleted);
1492 bin_width = bin_width < w ? w : bin_width;
1493 /* Display change counts aligned with "Bin" */
1494 number_width = 3;
1495 continue;
1496 }
1497
1498 if (max_change < change)
1499 max_change = change;
1500 }
1501 count = i; /* min(count, data->nr) */
1502
1503 /*
1504 * We have width = stat_width or term_columns() columns total.
1505 * We want a maximum of min(max_len, stat_name_width) for the name part.
1506 * We want a maximum of min(max_change, stat_graph_width) for the +- part.
1507 * We also need 1 for " " and 4 + decimal_width(max_change)
1508 * for " | NNNN " and one the empty column at the end, altogether
1509 * 6 + decimal_width(max_change).
1510 *
1511 * If there's not enough space, we will use the smaller of
1512 * stat_name_width (if set) and 5/8*width for the filename,
1513 * and the rest for constant elements + graph part, but no more
1514 * than stat_graph_width for the graph part.
1515 * (5/8 gives 50 for filename and 30 for the constant parts + graph
1516 * for the standard terminal size).
1517 *
1518 * In other words: stat_width limits the maximum width, and
1519 * stat_name_width fixes the maximum width of the filename,
1520 * and is also used to divide available columns if there
1521 * aren't enough.
1522 *
1523 * Binary files are displayed with "Bin XXX -> YYY bytes"
1524 * instead of the change count and graph. This part is treated
1525 * similarly to the graph part, except that it is not
1526 * "scaled". If total width is too small to accomodate the
1527 * guaranteed minimum width of the filename part and the
1528 * separators and this message, this message will "overflow"
1529 * making the line longer than the maximum width.
1530 */
1531
1532 if (options->stat_width == -1)
1533 width = term_columns();
1534 else
1535 width = options->stat_width ? options->stat_width : 80;
1536 number_width = decimal_width(max_change) > number_width ?
1537 decimal_width(max_change) : number_width;
1538
1539 if (options->stat_graph_width == -1)
1540 options->stat_graph_width = diff_stat_graph_width;
1541
1542 /*
1543 * Guarantee 3/8*16==6 for the graph part
1544 * and 5/8*16==10 for the filename part
1545 */
1546 if (width < 16 + 6 + number_width)
1547 width = 16 + 6 + number_width;
1548
1549 /*
1550 * First assign sizes that are wanted, ignoring available width.
1551 * strlen("Bin XXX -> YYY bytes") == bin_width, and the part
1552 * starting from "XXX" should fit in graph_width.
1553 */
1554 graph_width = max_change + 4 > bin_width ? max_change : bin_width - 4;
1555 if (options->stat_graph_width &&
1556 options->stat_graph_width < graph_width)
1557 graph_width = options->stat_graph_width;
1558
1559 name_width = (options->stat_name_width > 0 &&
1560 options->stat_name_width < max_len) ?
1561 options->stat_name_width : max_len;
1562
1563 /*
1564 * Adjust adjustable widths not to exceed maximum width
1565 */
1566 if (name_width + number_width + 6 + graph_width > width) {
1567 if (graph_width > width * 3/8 - number_width - 6)
1568 graph_width = width * 3/8 - number_width - 6;
1569 if (options->stat_graph_width &&
1570 graph_width > options->stat_graph_width)
1571 graph_width = options->stat_graph_width;
1572 if (name_width > width - number_width - 6 - graph_width)
1573 name_width = width - number_width - 6 - graph_width;
1574 else
1575 graph_width = width - number_width - 6 - name_width;
1576 }
1577
1578 /*
1579 * From here name_width is the width of the name area,
1580 * and graph_width is the width of the graph area.
1581 * max_change is used to scale graph properly.
1582 */
1583 for (i = 0; i < count; i++) {
1584 const char *prefix = "";
1585 char *name = data->files[i]->print_name;
1586 uintmax_t added = data->files[i]->added;
1587 uintmax_t deleted = data->files[i]->deleted;
1588 int name_len;
1589
1590 if (!data->files[i]->is_renamed &&
1591 (added + deleted == 0)) {
1592 total_files--;
1593 continue;
1594 }
1595 /*
1596 * "scale" the filename
1597 */
1598 len = name_width;
1599 name_len = strlen(name);
1600 if (name_width < name_len) {
1601 char *slash;
1602 prefix = "...";
1603 len -= 3;
1604 name += name_len - len;
1605 slash = strchr(name, '/');
1606 if (slash)
1607 name = slash;
1608 }
1609
1610 if (data->files[i]->is_binary) {
1611 fprintf(options->file, "%s", line_prefix);
1612 show_name(options->file, prefix, name, len);
1613 fprintf(options->file, " %*s ", number_width, "Bin");
1614 fprintf(options->file, "%s%"PRIuMAX"%s",
1615 del_c, deleted, reset);
1616 fprintf(options->file, " -> ");
1617 fprintf(options->file, "%s%"PRIuMAX"%s",
1618 add_c, added, reset);
1619 fprintf(options->file, " bytes");
1620 fprintf(options->file, "\n");
1621 continue;
1622 }
1623 else if (data->files[i]->is_unmerged) {
1624 fprintf(options->file, "%s", line_prefix);
1625 show_name(options->file, prefix, name, len);
1626 fprintf(options->file, " Unmerged\n");
1627 continue;
1628 }
1629
1630 /*
1631 * scale the add/delete
1632 */
1633 add = added;
1634 del = deleted;
1635 adds += add;
1636 dels += del;
1637
1638 if (graph_width <= max_change) {
1639 int total = add + del;
1640
1641 total = scale_linear(add + del, graph_width, max_change);
1642 if (total < 2 && add && del)
1643 /* width >= 2 due to the sanity check */
1644 total = 2;
1645 if (add < del) {
1646 add = scale_linear(add, graph_width, max_change);
1647 del = total - add;
1648 } else {
1649 del = scale_linear(del, graph_width, max_change);
1650 add = total - del;
1651 }
1652 }
1653 fprintf(options->file, "%s", line_prefix);
1654 show_name(options->file, prefix, name, len);
1655 fprintf(options->file, " %*"PRIuMAX"%s",
1656 number_width, added + deleted,
1657 added + deleted ? " " : "");
1658 show_graph(options->file, '+', add, add_c, reset);
1659 show_graph(options->file, '-', del, del_c, reset);
1660 fprintf(options->file, "\n");
1661 }
1662 for (i = count; i < data->nr; i++) {
1663 uintmax_t added = data->files[i]->added;
1664 uintmax_t deleted = data->files[i]->deleted;
1665 if (!data->files[i]->is_renamed &&
1666 (added + deleted == 0)) {
1667 total_files--;
1668 continue;
1669 }
1670 adds += added;
1671 dels += deleted;
1672 if (!extra_shown)
1673 fprintf(options->file, "%s ...\n", line_prefix);
1674 extra_shown = 1;
1675 }
1676 fprintf(options->file, "%s", line_prefix);
1677 print_stat_summary(options->file, total_files, adds, dels);
1678 }
1679
1680 static void show_shortstats(struct diffstat_t *data, struct diff_options *options)
1681 {
1682 int i, adds = 0, dels = 0, total_files = data->nr;
1683
1684 if (data->nr == 0)
1685 return;
1686
1687 for (i = 0; i < data->nr; i++) {
1688 if (!data->files[i]->is_binary &&
1689 !data->files[i]->is_unmerged) {
1690 int added = data->files[i]->added;
1691 int deleted= data->files[i]->deleted;
1692 if (!data->files[i]->is_renamed &&
1693 (added + deleted == 0)) {
1694 total_files--;
1695 } else {
1696 adds += added;
1697 dels += deleted;
1698 }
1699 }
1700 }
1701 if (options->output_prefix) {
1702 struct strbuf *msg = NULL;
1703 msg = options->output_prefix(options,
1704 options->output_prefix_data);
1705 fprintf(options->file, "%s", msg->buf);
1706 }
1707 print_stat_summary(options->file, total_files, adds, dels);
1708 }
1709
1710 static void show_numstat(struct diffstat_t *data, struct diff_options *options)
1711 {
1712 int i;
1713
1714 if (data->nr == 0)
1715 return;
1716
1717 for (i = 0; i < data->nr; i++) {
1718 struct diffstat_file *file = data->files[i];
1719
1720 if (options->output_prefix) {
1721 struct strbuf *msg = NULL;
1722 msg = options->output_prefix(options,
1723 options->output_prefix_data);
1724 fprintf(options->file, "%s", msg->buf);
1725 }
1726
1727 if (file->is_binary)
1728 fprintf(options->file, "-\t-\t");
1729 else
1730 fprintf(options->file,
1731 "%"PRIuMAX"\t%"PRIuMAX"\t",
1732 file->added, file->deleted);
1733 if (options->line_termination) {
1734 fill_print_name(file);
1735 if (!file->is_renamed)
1736 write_name_quoted(file->name, options->file,
1737 options->line_termination);
1738 else {
1739 fputs(file->print_name, options->file);
1740 putc(options->line_termination, options->file);
1741 }
1742 } else {
1743 if (file->is_renamed) {
1744 putc('\0', options->file);
1745 write_name_quoted(file->from_name, options->file, '\0');
1746 }
1747 write_name_quoted(file->name, options->file, '\0');
1748 }
1749 }
1750 }
1751
1752 struct dirstat_file {
1753 const char *name;
1754 unsigned long changed;
1755 };
1756
1757 struct dirstat_dir {
1758 struct dirstat_file *files;
1759 int alloc, nr, permille, cumulative;
1760 };
1761
1762 static long gather_dirstat(struct diff_options *opt, struct dirstat_dir *dir,
1763 unsigned long changed, const char *base, int baselen)
1764 {
1765 unsigned long this_dir = 0;
1766 unsigned int sources = 0;
1767 const char *line_prefix = "";
1768 struct strbuf *msg = NULL;
1769
1770 if (opt->output_prefix) {
1771 msg = opt->output_prefix(opt, opt->output_prefix_data);
1772 line_prefix = msg->buf;
1773 }
1774
1775 while (dir->nr) {
1776 struct dirstat_file *f = dir->files;
1777 int namelen = strlen(f->name);
1778 unsigned long this;
1779 char *slash;
1780
1781 if (namelen < baselen)
1782 break;
1783 if (memcmp(f->name, base, baselen))
1784 break;
1785 slash = strchr(f->name + baselen, '/');
1786 if (slash) {
1787 int newbaselen = slash + 1 - f->name;
1788 this = gather_dirstat(opt, dir, changed, f->name, newbaselen);
1789 sources++;
1790 } else {
1791 this = f->changed;
1792 dir->files++;
1793 dir->nr--;
1794 sources += 2;
1795 }
1796 this_dir += this;
1797 }
1798
1799 /*
1800 * We don't report dirstat's for
1801 * - the top level
1802 * - or cases where everything came from a single directory
1803 * under this directory (sources == 1).
1804 */
1805 if (baselen && sources != 1) {
1806 if (this_dir) {
1807 int permille = this_dir * 1000 / changed;
1808 if (permille >= dir->permille) {
1809 fprintf(opt->file, "%s%4d.%01d%% %.*s\n", line_prefix,
1810 permille / 10, permille % 10, baselen, base);
1811 if (!dir->cumulative)
1812 return 0;
1813 }
1814 }
1815 }
1816 return this_dir;
1817 }
1818
1819 static int dirstat_compare(const void *_a, const void *_b)
1820 {
1821 const struct dirstat_file *a = _a;
1822 const struct dirstat_file *b = _b;
1823 return strcmp(a->name, b->name);
1824 }
1825
1826 static void show_dirstat(struct diff_options *options)
1827 {
1828 int i;
1829 unsigned long changed;
1830 struct dirstat_dir dir;
1831 struct diff_queue_struct *q = &diff_queued_diff;
1832
1833 dir.files = NULL;
1834 dir.alloc = 0;
1835 dir.nr = 0;
1836 dir.permille = options->dirstat_permille;
1837 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1838
1839 changed = 0;
1840 for (i = 0; i < q->nr; i++) {
1841 struct diff_filepair *p = q->queue[i];
1842 const char *name;
1843 unsigned long copied, added, damage;
1844 int content_changed;
1845
1846 name = p->two->path ? p->two->path : p->one->path;
1847
1848 if (p->one->sha1_valid && p->two->sha1_valid)
1849 content_changed = hashcmp(p->one->sha1, p->two->sha1);
1850 else
1851 content_changed = 1;
1852
1853 if (!content_changed) {
1854 /*
1855 * The SHA1 has not changed, so pre-/post-content is
1856 * identical. We can therefore skip looking at the
1857 * file contents altogether.
1858 */
1859 damage = 0;
1860 goto found_damage;
1861 }
1862
1863 if (DIFF_OPT_TST(options, DIRSTAT_BY_FILE)) {
1864 /*
1865 * In --dirstat-by-file mode, we don't really need to
1866 * look at the actual file contents at all.
1867 * The fact that the SHA1 changed is enough for us to
1868 * add this file to the list of results
1869 * (with each file contributing equal damage).
1870 */
1871 damage = 1;
1872 goto found_damage;
1873 }
1874
1875 if (DIFF_FILE_VALID(p->one) && DIFF_FILE_VALID(p->two)) {
1876 diff_populate_filespec(p->one, 0);
1877 diff_populate_filespec(p->two, 0);
1878 diffcore_count_changes(p->one, p->two, NULL, NULL, 0,
1879 &copied, &added);
1880 diff_free_filespec_data(p->one);
1881 diff_free_filespec_data(p->two);
1882 } else if (DIFF_FILE_VALID(p->one)) {
1883 diff_populate_filespec(p->one, 1);
1884 copied = added = 0;
1885 diff_free_filespec_data(p->one);
1886 } else if (DIFF_FILE_VALID(p->two)) {
1887 diff_populate_filespec(p->two, 1);
1888 copied = 0;
1889 added = p->two->size;
1890 diff_free_filespec_data(p->two);
1891 } else
1892 continue;
1893
1894 /*
1895 * Original minus copied is the removed material,
1896 * added is the new material. They are both damages
1897 * made to the preimage.
1898 * If the resulting damage is zero, we know that
1899 * diffcore_count_changes() considers the two entries to
1900 * be identical, but since content_changed is true, we
1901 * know that there must have been _some_ kind of change,
1902 * so we force all entries to have damage > 0.
1903 */
1904 damage = (p->one->size - copied) + added;
1905 if (!damage)
1906 damage = 1;
1907
1908 found_damage:
1909 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1910 dir.files[dir.nr].name = name;
1911 dir.files[dir.nr].changed = damage;
1912 changed += damage;
1913 dir.nr++;
1914 }
1915
1916 /* This can happen even with many files, if everything was renames */
1917 if (!changed)
1918 return;
1919
1920 /* Show all directories with more than x% of the changes */
1921 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1922 gather_dirstat(options, &dir, changed, "", 0);
1923 }
1924
1925 static void show_dirstat_by_line(struct diffstat_t *data, struct diff_options *options)
1926 {
1927 int i;
1928 unsigned long changed;
1929 struct dirstat_dir dir;
1930
1931 if (data->nr == 0)
1932 return;
1933
1934 dir.files = NULL;
1935 dir.alloc = 0;
1936 dir.nr = 0;
1937 dir.permille = options->dirstat_permille;
1938 dir.cumulative = DIFF_OPT_TST(options, DIRSTAT_CUMULATIVE);
1939
1940 changed = 0;
1941 for (i = 0; i < data->nr; i++) {
1942 struct diffstat_file *file = data->files[i];
1943 unsigned long damage = file->added + file->deleted;
1944 if (file->is_binary)
1945 /*
1946 * binary files counts bytes, not lines. Must find some
1947 * way to normalize binary bytes vs. textual lines.
1948 * The following heuristic assumes that there are 64
1949 * bytes per "line".
1950 * This is stupid and ugly, but very cheap...
1951 */
1952 damage = (damage + 63) / 64;
1953 ALLOC_GROW(dir.files, dir.nr + 1, dir.alloc);
1954 dir.files[dir.nr].name = file->name;
1955 dir.files[dir.nr].changed = damage;
1956 changed += damage;
1957 dir.nr++;
1958 }
1959
1960 /* This can happen even with many files, if everything was renames */
1961 if (!changed)
1962 return;
1963
1964 /* Show all directories with more than x% of the changes */
1965 qsort(dir.files, dir.nr, sizeof(dir.files[0]), dirstat_compare);
1966 gather_dirstat(options, &dir, changed, "", 0);
1967 }
1968
1969 static void free_diffstat_info(struct diffstat_t *diffstat)
1970 {
1971 int i;
1972 for (i = 0; i < diffstat->nr; i++) {
1973 struct diffstat_file *f = diffstat->files[i];
1974 if (f->name != f->print_name)
1975 free(f->print_name);
1976 free(f->name);
1977 free(f->from_name);
1978 free(f);
1979 }
1980 free(diffstat->files);
1981 }
1982
1983 struct checkdiff_t {
1984 const char *filename;
1985 int lineno;
1986 int conflict_marker_size;
1987 struct diff_options *o;
1988 unsigned ws_rule;
1989 unsigned status;
1990 };
1991
1992 static int is_conflict_marker(const char *line, int marker_size, unsigned long len)
1993 {
1994 char firstchar;
1995 int cnt;
1996
1997 if (len < marker_size + 1)
1998 return 0;
1999 firstchar = line[0];
2000 switch (firstchar) {
2001 case '=': case '>': case '<': case '|':
2002 break;
2003 default:
2004 return 0;
2005 }
2006 for (cnt = 1; cnt < marker_size; cnt++)
2007 if (line[cnt] != firstchar)
2008 return 0;
2009 /* line[1] thru line[marker_size-1] are same as firstchar */
2010 if (len < marker_size + 1 || !isspace(line[marker_size]))
2011 return 0;
2012 return 1;
2013 }
2014
2015 static void checkdiff_consume(void *priv, char *line, unsigned long len)
2016 {
2017 struct checkdiff_t *data = priv;
2018 int marker_size = data->conflict_marker_size;
2019 const char *ws = diff_get_color(data->o->use_color, DIFF_WHITESPACE);
2020 const char *reset = diff_get_color(data->o->use_color, DIFF_RESET);
2021 const char *set = diff_get_color(data->o->use_color, DIFF_FILE_NEW);
2022 char *err;
2023 char *line_prefix = "";
2024 struct strbuf *msgbuf;
2025
2026 assert(data->o);
2027 if (data->o->output_prefix) {
2028 msgbuf = data->o->output_prefix(data->o,
2029 data->o->output_prefix_data);
2030 line_prefix = msgbuf->buf;
2031 }
2032
2033 if (line[0] == '+') {
2034 unsigned bad;
2035 data->lineno++;
2036 if (is_conflict_marker(line + 1, marker_size, len - 1)) {
2037 data->status |= 1;
2038 fprintf(data->o->file,
2039 "%s%s:%d: leftover conflict marker\n",
2040 line_prefix, data->filename, data->lineno);
2041 }
2042 bad = ws_check(line + 1, len - 1, data->ws_rule);
2043 if (!bad)
2044 return;
2045 data->status |= bad;
2046 err = whitespace_error_string(bad);
2047 fprintf(data->o->file, "%s%s:%d: %s.\n",
2048 line_prefix, data->filename, data->lineno, err);
2049 free(err);
2050 emit_line(data->o, set, reset, line, 1);
2051 ws_check_emit(line + 1, len - 1, data->ws_rule,
2052 data->o->file, set, reset, ws);
2053 } else if (line[0] == ' ') {
2054 data->lineno++;
2055 } else if (line[0] == '@') {
2056 char *plus = strchr(line, '+');
2057 if (plus)
2058 data->lineno = strtol(plus, NULL, 10) - 1;
2059 else
2060 die("invalid diff");
2061 }
2062 }
2063
2064 static unsigned char *deflate_it(char *data,
2065 unsigned long size,
2066 unsigned long *result_size)
2067 {
2068 int bound;
2069 unsigned char *deflated;
2070 git_zstream stream;
2071
2072 memset(&stream, 0, sizeof(stream));
2073 git_deflate_init(&stream, zlib_compression_level);
2074 bound = git_deflate_bound(&stream, size);
2075 deflated = xmalloc(bound);
2076 stream.next_out = deflated;
2077 stream.avail_out = bound;
2078
2079 stream.next_in = (unsigned char *)data;
2080 stream.avail_in = size;
2081 while (git_deflate(&stream, Z_FINISH) == Z_OK)
2082 ; /* nothing */
2083 git_deflate_end(&stream);
2084 *result_size = stream.total_out;
2085 return deflated;
2086 }
2087
2088 static void emit_binary_diff_body(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
2089 {
2090 void *cp;
2091 void *delta;
2092 void *deflated;
2093 void *data;
2094 unsigned long orig_size;
2095 unsigned long delta_size;
2096 unsigned long deflate_size;
2097 unsigned long data_size;
2098
2099 /* We could do deflated delta, or we could do just deflated two,
2100 * whichever is smaller.
2101 */
2102 delta = NULL;
2103 deflated = deflate_it(two->ptr, two->size, &deflate_size);
2104 if (one->size && two->size) {
2105 delta = diff_delta(one->ptr, one->size,
2106 two->ptr, two->size,
2107 &delta_size, deflate_size);
2108 if (delta) {
2109 void *to_free = delta;
2110 orig_size = delta_size;
2111 delta = deflate_it(delta, delta_size, &delta_size);
2112 free(to_free);
2113 }
2114 }
2115
2116 if (delta && delta_size < deflate_size) {
2117 fprintf(file, "%sdelta %lu\n", prefix, orig_size);
2118 free(deflated);
2119 data = delta;
2120 data_size = delta_size;
2121 }
2122 else {
2123 fprintf(file, "%sliteral %lu\n", prefix, two->size);
2124 free(delta);
2125 data = deflated;
2126 data_size = deflate_size;
2127 }
2128
2129 /* emit data encoded in base85 */
2130 cp = data;
2131 while (data_size) {
2132 int bytes = (52 < data_size) ? 52 : data_size;
2133 char line[70];
2134 data_size -= bytes;
2135 if (bytes <= 26)
2136 line[0] = bytes + 'A' - 1;
2137 else
2138 line[0] = bytes - 26 + 'a' - 1;
2139 encode_85(line + 1, cp, bytes);
2140 cp = (char *) cp + bytes;
2141 fprintf(file, "%s", prefix);
2142 fputs(line, file);
2143 fputc('\n', file);
2144 }
2145 fprintf(file, "%s\n", prefix);
2146 free(data);
2147 }
2148
2149 static void emit_binary_diff(FILE *file, mmfile_t *one, mmfile_t *two, char *prefix)
2150 {
2151 fprintf(file, "%sGIT binary patch\n", prefix);
2152 emit_binary_diff_body(file, one, two, prefix);
2153 emit_binary_diff_body(file, two, one, prefix);
2154 }
2155
2156 int diff_filespec_is_binary(struct diff_filespec *one)
2157 {
2158 if (one->is_binary == -1) {
2159 diff_filespec_load_driver(one);
2160 if (one->driver->binary != -1)
2161 one->is_binary = one->driver->binary;
2162 else {
2163 if (!one->data && DIFF_FILE_VALID(one))
2164 diff_populate_filespec(one, 0);
2165 if (one->data)
2166 one->is_binary = buffer_is_binary(one->data,
2167 one->size);
2168 if (one->is_binary == -1)
2169 one->is_binary = 0;
2170 }
2171 }
2172 return one->is_binary;
2173 }
2174
2175 static const struct userdiff_funcname *diff_funcname_pattern(struct diff_filespec *one)
2176 {
2177 diff_filespec_load_driver(one);
2178 return one->driver->funcname.pattern ? &one->driver->funcname : NULL;
2179 }
2180
2181 void diff_set_mnemonic_prefix(struct diff_options *options, const char *a, const char *b)
2182 {
2183 if (!options->a_prefix)
2184 options->a_prefix = a;
2185 if (!options->b_prefix)
2186 options->b_prefix = b;
2187 }
2188
2189 struct userdiff_driver *get_textconv(struct diff_filespec *one)
2190 {
2191 if (!DIFF_FILE_VALID(one))
2192 return NULL;
2193
2194 diff_filespec_load_driver(one);
2195 return userdiff_get_textconv(one->driver);
2196 }
2197
2198 static void builtin_diff(const char *name_a,
2199 const char *name_b,
2200 struct diff_filespec *one,
2201 struct diff_filespec *two,
2202 const char *xfrm_msg,
2203 int must_show_header,
2204 struct diff_options *o,
2205 int complete_rewrite)
2206 {
2207 mmfile_t mf1, mf2;
2208 const char *lbl[2];
2209 char *a_one, *b_two;
2210 const char *set = diff_get_color_opt(o, DIFF_METAINFO);
2211 const char *reset = diff_get_color_opt(o, DIFF_RESET);
2212 const char *a_prefix, *b_prefix;
2213 struct userdiff_driver *textconv_one = NULL;
2214 struct userdiff_driver *textconv_two = NULL;
2215 struct strbuf header = STRBUF_INIT;
2216 struct strbuf *msgbuf;
2217 char *line_prefix = "";
2218
2219 if (o->output_prefix) {
2220 msgbuf = o->output_prefix(o, o->output_prefix_data);
2221 line_prefix = msgbuf->buf;
2222 }
2223
2224 if (DIFF_OPT_TST(o, SUBMODULE_LOG) &&
2225 (!one->mode || S_ISGITLINK(one->mode)) &&
2226 (!two->mode || S_ISGITLINK(two->mode))) {
2227 const char *del = diff_get_color_opt(o, DIFF_FILE_OLD);
2228 const char *add = diff_get_color_opt(o, DIFF_FILE_NEW);
2229 show_submodule_summary(o->file, one ? one->path : two->path,
2230 one->sha1, two->sha1, two->dirty_submodule,
2231 del, add, reset);
2232 return;
2233 }
2234
2235 if (DIFF_OPT_TST(o, ALLOW_TEXTCONV)) {
2236 textconv_one = get_textconv(one);
2237 textconv_two = get_textconv(two);
2238 }
2239
2240 diff_set_mnemonic_prefix(o, "a/", "b/");
2241 if (DIFF_OPT_TST(o, REVERSE_DIFF)) {
2242 a_prefix = o->b_prefix;
2243 b_prefix = o->a_prefix;
2244 } else {
2245 a_prefix = o->a_prefix;
2246 b_prefix = o->b_prefix;
2247 }
2248
2249 /* Never use a non-valid filename anywhere if at all possible */
2250 name_a = DIFF_FILE_VALID(one) ? name_a : name_b;
2251 name_b = DIFF_FILE_VALID(two) ? name_b : name_a;
2252
2253 a_one = quote_two(a_prefix, name_a + (*name_a == '/'));
2254 b_two = quote_two(b_prefix, name_b + (*name_b == '/'));
2255 lbl[0] = DIFF_FILE_VALID(one) ? a_one : "/dev/null";
2256 lbl[1] = DIFF_FILE_VALID(two) ? b_two : "/dev/null";
2257 strbuf_addf(&header, "%s%sdiff --git %s %s%s\n", line_prefix, set, a_one, b_two, reset);
2258 if (lbl[0][0] == '/') {
2259 /* /dev/null */
2260 strbuf_addf(&header, "%s%snew file mode %06o%s\n", line_prefix, set, two->mode, reset);
2261 if (xfrm_msg)
2262 strbuf_addstr(&header, xfrm_msg);
2263 must_show_header = 1;
2264 }
2265 else if (lbl[1][0] == '/') {
2266 strbuf_addf(&header, "%s%sdeleted file mode %06o%s\n", line_prefix, set, one->mode, reset);
2267 if (xfrm_msg)
2268 strbuf_addstr(&header, xfrm_msg);
2269 must_show_header = 1;
2270 }
2271 else {
2272 if (one->mode != two->mode) {
2273 strbuf_addf(&header, "%s%sold mode %06o%s\n", line_prefix, set, one->mode, reset);
2274 strbuf_addf(&header, "%s%snew mode %06o%s\n", line_prefix, set, two->mode, reset);
2275 must_show_header = 1;
2276 }
2277 if (xfrm_msg)
2278 strbuf_addstr(&header, xfrm_msg);
2279
2280 /*
2281 * we do not run diff between different kind
2282 * of objects.
2283 */
2284 if ((one->mode ^ two->mode) & S_IFMT)
2285 goto free_ab_and_return;
2286 if (complete_rewrite &&
2287 (textconv_one || !diff_filespec_is_binary(one)) &&
2288 (textconv_two || !diff_filespec_is_binary(two))) {
2289 fprintf(o->file, "%s", header.buf);
2290 strbuf_reset(&header);
2291 emit_rewrite_diff(name_a, name_b, one, two,
2292 textconv_one, textconv_two, o);
2293 o->found_changes = 1;
2294 goto free_ab_and_return;
2295 }
2296 }
2297
2298 if (o->irreversible_delete && lbl[1][0] == '/') {
2299 fprintf(o->file, "%s", header.buf);
2300 strbuf_reset(&header);
2301 goto free_ab_and_return;
2302 } else if (!DIFF_OPT_TST(o, TEXT) &&
2303 ( (!textconv_one && diff_filespec_is_binary(one)) ||
2304 (!textconv_two && diff_filespec_is_binary(two)) )) {
2305 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2306 die("unable to read files to diff");
2307 /* Quite common confusing case */
2308 if (mf1.size == mf2.size &&
2309 !memcmp(mf1.ptr, mf2.ptr, mf1.size)) {
2310 if (must_show_header)
2311 fprintf(o->file, "%s", header.buf);
2312 goto free_ab_and_return;
2313 }
2314 fprintf(o->file, "%s", header.buf);
2315 strbuf_reset(&header);
2316 if (DIFF_OPT_TST(o, BINARY))
2317 emit_binary_diff(o->file, &mf1, &mf2, line_prefix);
2318 else
2319 fprintf(o->file, "%sBinary files %s and %s differ\n",
2320 line_prefix, lbl[0], lbl[1]);
2321 o->found_changes = 1;
2322 } else {
2323 /* Crazy xdl interfaces.. */
2324 const char *diffopts = getenv("GIT_DIFF_OPTS");
2325 xpparam_t xpp;
2326 xdemitconf_t xecfg;
2327 struct emit_callback ecbdata;
2328 const struct userdiff_funcname *pe;
2329
2330 if (must_show_header) {
2331 fprintf(o->file, "%s", header.buf);
2332 strbuf_reset(&header);
2333 }
2334
2335 mf1.size = fill_textconv(textconv_one, one, &mf1.ptr);
2336 mf2.size = fill_textconv(textconv_two, two, &mf2.ptr);
2337
2338 pe = diff_funcname_pattern(one);
2339 if (!pe)
2340 pe = diff_funcname_pattern(two);
2341
2342 memset(&xpp, 0, sizeof(xpp));
2343 memset(&xecfg, 0, sizeof(xecfg));
2344 memset(&ecbdata, 0, sizeof(ecbdata));
2345 ecbdata.label_path = lbl;
2346 ecbdata.color_diff = want_color(o->use_color);
2347 ecbdata.found_changesp = &o->found_changes;
2348 ecbdata.ws_rule = whitespace_rule(name_b ? name_b : name_a);
2349 if (ecbdata.ws_rule & WS_BLANK_AT_EOF)
2350 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2351 ecbdata.opt = o;
2352 ecbdata.header = header.len ? &header : NULL;
2353 xpp.flags = o->xdl_opts;
2354 xecfg.ctxlen = o->context;
2355 xecfg.interhunkctxlen = o->interhunkcontext;
2356 xecfg.flags = XDL_EMIT_FUNCNAMES;
2357 if (DIFF_OPT_TST(o, FUNCCONTEXT))
2358 xecfg.flags |= XDL_EMIT_FUNCCONTEXT;
2359 if (pe)
2360 xdiff_set_find_func(&xecfg, pe->pattern, pe->cflags);
2361 if (!diffopts)
2362 ;
2363 else if (!prefixcmp(diffopts, "--unified="))
2364 xecfg.ctxlen = strtoul(diffopts + 10, NULL, 10);
2365 else if (!prefixcmp(diffopts, "-u"))
2366 xecfg.ctxlen = strtoul(diffopts + 2, NULL, 10);
2367 if (o->word_diff)
2368 init_diff_words_data(&ecbdata, o, one, two);
2369 xdi_diff_outf(&mf1, &mf2, fn_out_consume, &ecbdata,
2370 &xpp, &xecfg);
2371 if (o->word_diff)
2372 free_diff_words_data(&ecbdata);
2373 if (textconv_one)
2374 free(mf1.ptr);
2375 if (textconv_two)
2376 free(mf2.ptr);
2377 xdiff_clear_find_func(&xecfg);
2378 }
2379
2380 free_ab_and_return:
2381 strbuf_release(&header);
2382 diff_free_filespec_data(one);
2383 diff_free_filespec_data(two);
2384 free(a_one);
2385 free(b_two);
2386 return;
2387 }
2388
2389 static void builtin_diffstat(const char *name_a, const char *name_b,
2390 struct diff_filespec *one,
2391 struct diff_filespec *two,
2392 struct diffstat_t *diffstat,
2393 struct diff_options *o,
2394 int complete_rewrite)
2395 {
2396 mmfile_t mf1, mf2;
2397 struct diffstat_file *data;
2398
2399 data = diffstat_add(diffstat, name_a, name_b);
2400
2401 if (!one || !two) {
2402 data->is_unmerged = 1;
2403 return;
2404 }
2405
2406 if (diff_filespec_is_binary(one) || diff_filespec_is_binary(two)) {
2407 data->is_binary = 1;
2408 data->added = diff_filespec_size(two);
2409 data->deleted = diff_filespec_size(one);
2410 }
2411
2412 else if (complete_rewrite) {
2413 diff_populate_filespec(one, 0);
2414 diff_populate_filespec(two, 0);
2415 data->deleted = count_lines(one->data, one->size);
2416 data->added = count_lines(two->data, two->size);
2417 }
2418
2419 else {
2420 /* Crazy xdl interfaces.. */
2421 xpparam_t xpp;
2422 xdemitconf_t xecfg;
2423
2424 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2425 die("unable to read files to diff");
2426
2427 memset(&xpp, 0, sizeof(xpp));
2428 memset(&xecfg, 0, sizeof(xecfg));
2429 xpp.flags = o->xdl_opts;
2430 xecfg.ctxlen = o->context;
2431 xecfg.interhunkctxlen = o->interhunkcontext;
2432 xdi_diff_outf(&mf1, &mf2, diffstat_consume, diffstat,
2433 &xpp, &xecfg);
2434 }
2435
2436 diff_free_filespec_data(one);
2437 diff_free_filespec_data(two);
2438 }
2439
2440 static void builtin_checkdiff(const char *name_a, const char *name_b,
2441 const char *attr_path,
2442 struct diff_filespec *one,
2443 struct diff_filespec *two,
2444 struct diff_options *o)
2445 {
2446 mmfile_t mf1, mf2;
2447 struct checkdiff_t data;
2448
2449 if (!two)
2450 return;
2451
2452 memset(&data, 0, sizeof(data));
2453 data.filename = name_b ? name_b : name_a;
2454 data.lineno = 0;
2455 data.o = o;
2456 data.ws_rule = whitespace_rule(attr_path);
2457 data.conflict_marker_size = ll_merge_marker_size(attr_path);
2458
2459 if (fill_mmfile(&mf1, one) < 0 || fill_mmfile(&mf2, two) < 0)
2460 die("unable to read files to diff");
2461
2462 /*
2463 * All the other codepaths check both sides, but not checking
2464 * the "old" side here is deliberate. We are checking the newly
2465 * introduced changes, and as long as the "new" side is text, we
2466 * can and should check what it introduces.
2467 */
2468 if (diff_filespec_is_binary(two))
2469 goto free_and_return;
2470 else {
2471 /* Crazy xdl interfaces.. */
2472 xpparam_t xpp;
2473 xdemitconf_t xecfg;
2474
2475 memset(&xpp, 0, sizeof(xpp));
2476 memset(&xecfg, 0, sizeof(xecfg));
2477 xecfg.ctxlen = 1; /* at least one context line */
2478 xpp.flags = 0;
2479 xdi_diff_outf(&mf1, &mf2, checkdiff_consume, &data,
2480 &xpp, &xecfg);
2481
2482 if (data.ws_rule & WS_BLANK_AT_EOF) {
2483 struct emit_callback ecbdata;
2484 int blank_at_eof;
2485
2486 ecbdata.ws_rule = data.ws_rule;
2487 check_blank_at_eof(&mf1, &mf2, &ecbdata);
2488 blank_at_eof = ecbdata.blank_at_eof_in_postimage;
2489
2490 if (blank_at_eof) {
2491 static char *err;
2492 if (!err)
2493 err = whitespace_error_string(WS_BLANK_AT_EOF);
2494 fprintf(o->file, "%s:%d: %s.\n",
2495 data.filename, blank_at_eof, err);
2496 data.status = 1; /* report errors */
2497 }
2498 }
2499 }
2500 free_and_return:
2501 diff_free_filespec_data(one);
2502 diff_free_filespec_data(two);
2503 if (data.status)
2504 DIFF_OPT_SET(o, CHECK_FAILED);
2505 }
2506
2507 struct diff_filespec *alloc_filespec(const char *path)
2508 {
2509 int namelen = strlen(path);
2510 struct diff_filespec *spec = xmalloc(sizeof(*spec) + namelen + 1);
2511
2512 memset(spec, 0, sizeof(*spec));
2513 spec->path = (char *)(spec + 1);
2514 memcpy(spec->path, path, namelen+1);
2515 spec->count = 1;
2516 spec->is_binary = -1;
2517 return spec;
2518 }
2519
2520 void free_filespec(struct diff_filespec *spec)
2521 {
2522 if (!--spec->count) {
2523 diff_free_filespec_data(spec);
2524 free(spec);
2525 }
2526 }
2527
2528 void fill_filespec(struct diff_filespec *spec, const unsigned char *sha1,
2529 unsigned short mode)
2530 {
2531 if (mode) {
2532 spec->mode = canon_mode(mode);
2533 hashcpy(spec->sha1, sha1);
2534 spec->sha1_valid = !is_null_sha1(sha1);
2535 }
2536 }
2537
2538 /*
2539 * Given a name and sha1 pair, if the index tells us the file in
2540 * the work tree has that object contents, return true, so that
2541 * prepare_temp_file() does not have to inflate and extract.
2542 */
2543 static int reuse_worktree_file(const char *name, const unsigned char *sha1, int want_file)
2544 {
2545 struct cache_entry *ce;
2546 struct stat st;
2547 int pos, len;
2548
2549 /*
2550 * We do not read the cache ourselves here, because the
2551 * benchmark with my previous version that always reads cache
2552 * shows that it makes things worse for diff-tree comparing
2553 * two linux-2.6 kernel trees in an already checked out work
2554 * tree. This is because most diff-tree comparisons deal with
2555 * only a small number of files, while reading the cache is
2556 * expensive for a large project, and its cost outweighs the
2557 * savings we get by not inflating the object to a temporary
2558 * file. Practically, this code only helps when we are used
2559 * by diff-cache --cached, which does read the cache before
2560 * calling us.
2561 */
2562 if (!active_cache)
2563 return 0;
2564
2565 /* We want to avoid the working directory if our caller
2566 * doesn't need the data in a normal file, this system
2567 * is rather slow with its stat/open/mmap/close syscalls,
2568 * and the object is contained in a pack file. The pack
2569 * is probably already open and will be faster to obtain
2570 * the data through than the working directory. Loose
2571 * objects however would tend to be slower as they need
2572 * to be individually opened and inflated.
2573 */
2574 if (!FAST_WORKING_DIRECTORY && !want_file && has_sha1_pack(sha1))
2575 return 0;
2576
2577 len = strlen(name);
2578 pos = cache_name_pos(name, len);
2579 if (pos < 0)
2580 return 0;
2581 ce = active_cache[pos];
2582
2583 /*
2584 * This is not the sha1 we are looking for, or
2585 * unreusable because it is not a regular file.
2586 */
2587 if (hashcmp(sha1, ce->sha1) || !S_ISREG(ce->ce_mode))
2588 return 0;
2589
2590 /*
2591 * If ce is marked as "assume unchanged", there is no
2592 * guarantee that work tree matches what we are looking for.
2593 */
2594 if ((ce->ce_flags & CE_VALID) || ce_skip_worktree(ce))
2595 return 0;
2596
2597 /*
2598 * If ce matches the file in the work tree, we can reuse it.
2599 */
2600 if (ce_uptodate(ce) ||
2601 (!lstat(name, &st) && !ce_match_stat(ce, &st, 0)))
2602 return 1;
2603
2604 return 0;
2605 }
2606
2607 static int populate_from_stdin(struct diff_filespec *s)
2608 {
2609 struct strbuf buf = STRBUF_INIT;
2610 size_t size = 0;
2611
2612 if (strbuf_read(&buf, 0, 0) < 0)
2613 return error("error while reading from stdin %s",
2614 strerror(errno));
2615
2616 s->should_munmap = 0;
2617 s->data = strbuf_detach(&buf, &size);
2618 s->size = size;
2619 s->should_free = 1;
2620 return 0;
2621 }
2622
2623 static int diff_populate_gitlink(struct diff_filespec *s, int size_only)
2624 {
2625 int len;
2626 char *data = xmalloc(100), *dirty = "";
2627
2628 /* Are we looking at the work tree? */
2629 if (s->dirty_submodule)
2630 dirty = "-dirty";
2631
2632 len = snprintf(data, 100,
2633 "Subproject commit %s%s\n", sha1_to_hex(s->sha1), dirty);
2634 s->data = data;
2635 s->size = len;
2636 s->should_free = 1;
2637 if (size_only) {
2638 s->data = NULL;
2639 free(data);
2640 }
2641 return 0;
2642 }
2643
2644 /*
2645 * While doing rename detection and pickaxe operation, we may need to
2646 * grab the data for the blob (or file) for our own in-core comparison.
2647 * diff_filespec has data and size fields for this purpose.
2648 */
2649 int diff_populate_filespec(struct diff_filespec *s, int size_only)
2650 {
2651 int err = 0;
2652 if (!DIFF_FILE_VALID(s))
2653 die("internal error: asking to populate invalid file.");
2654 if (S_ISDIR(s->mode))
2655 return -1;
2656
2657 if (s->data)
2658 return 0;
2659
2660 if (size_only && 0 < s->size)
2661 return 0;
2662
2663 if (S_ISGITLINK(s->mode))
2664 return diff_populate_gitlink(s, size_only);
2665
2666 if (!s->sha1_valid ||
2667 reuse_worktree_file(s->path, s->sha1, 0)) {
2668 struct strbuf buf = STRBUF_INIT;
2669 struct stat st;
2670 int fd;
2671
2672 if (!strcmp(s->path, "-"))
2673 return populate_from_stdin(s);
2674
2675 if (lstat(s->path, &st) < 0) {
2676 if (errno == ENOENT) {
2677 err_empty:
2678 err = -1;
2679 empty:
2680 s->data = (char *)"";
2681 s->size = 0;
2682 return err;
2683 }
2684 }
2685 s->size = xsize_t(st.st_size);
2686 if (!s->size)
2687 goto empty;
2688 if (S_ISLNK(st.st_mode)) {
2689 struct strbuf sb = STRBUF_INIT;
2690
2691 if (strbuf_readlink(&sb, s->path, s->size))
2692 goto err_empty;
2693 s->size = sb.len;
2694 s->data = strbuf_detach(&sb, NULL);
2695 s->should_free = 1;
2696 return 0;
2697 }
2698 if (size_only)
2699 return 0;
2700 fd = open(s->path, O_RDONLY);
2701 if (fd < 0)
2702 goto err_empty;
2703 s->data = xmmap(NULL, s->size, PROT_READ, MAP_PRIVATE, fd, 0);
2704 close(fd);
2705 s->should_munmap = 1;
2706
2707 /*
2708 * Convert from working tree format to canonical git format
2709 */
2710 if (convert_to_git(s->path, s->data, s->size, &buf, safe_crlf)) {
2711 size_t size = 0;
2712 munmap(s->data, s->size);
2713 s->should_munmap = 0;
2714 s->data = strbuf_detach(&buf, &size);
2715 s->size = size;
2716 s->should_free = 1;
2717 }
2718 }
2719 else {
2720 enum object_type type;
2721 if (size_only) {
2722 type = sha1_object_info(s->sha1, &s->size);
2723 if (type < 0)
2724 die("unable to read %s", sha1_to_hex(s->sha1));
2725 } else {
2726 s->data = read_sha1_file(s->sha1, &type, &s->size);
2727 if (!s->data)
2728 die("unable to read %s", sha1_to_hex(s->sha1));
2729 s->should_free = 1;
2730 }
2731 }
2732 return 0;
2733 }
2734
2735 void diff_free_filespec_blob(struct diff_filespec *s)
2736 {
2737 if (s->should_free)
2738 free(s->data);
2739 else if (s->should_munmap)
2740 munmap(s->data, s->size);
2741
2742 if (s->should_free || s->should_munmap) {
2743 s->should_free = s->should_munmap = 0;
2744 s->data = NULL;
2745 }
2746 }
2747
2748 void diff_free_filespec_data(struct diff_filespec *s)
2749 {
2750 diff_free_filespec_blob(s);
2751 free(s->cnt_data);
2752 s->cnt_data = NULL;
2753 }
2754
2755 static void prep_temp_blob(const char *path, struct diff_tempfile *temp,
2756 void *blob,
2757 unsigned long size,
2758 const unsigned char *sha1,
2759 int mode)
2760 {
2761 int fd;
2762 struct strbuf buf = STRBUF_INIT;
2763 struct strbuf template = STRBUF_INIT;
2764 char *path_dup = xstrdup(path);
2765 const char *base = basename(path_dup);
2766
2767 /* Generate "XXXXXX_basename.ext" */
2768 strbuf_addstr(&template, "XXXXXX_");
2769 strbuf_addstr(&template, base);
2770
2771 fd = git_mkstemps(temp->tmp_path, PATH_MAX, template.buf,
2772 strlen(base) + 1);
2773 if (fd < 0)
2774 die_errno("unable to create temp-file");
2775 if (convert_to_working_tree(path,
2776 (const char *)blob, (size_t)size, &buf)) {
2777 blob = buf.buf;
2778 size = buf.len;
2779 }
2780 if (write_in_full(fd, blob, size) != size)
2781 die_errno("unable to write temp-file");
2782 close(fd);
2783 temp->name = temp->tmp_path;
2784 strcpy(temp->hex, sha1_to_hex(sha1));
2785 temp->hex[40] = 0;
2786 sprintf(temp->mode, "%06o", mode);
2787 strbuf_release(&buf);
2788 strbuf_release(&template);
2789 free(path_dup);
2790 }
2791
2792 static struct diff_tempfile *prepare_temp_file(const char *name,
2793 struct diff_filespec *one)
2794 {
2795 struct diff_tempfile *temp = claim_diff_tempfile();
2796
2797 if (!DIFF_FILE_VALID(one)) {
2798 not_a_valid_file:
2799 /* A '-' entry produces this for file-2, and
2800 * a '+' entry produces this for file-1.
2801 */
2802 temp->name = "/dev/null";
2803 strcpy(temp->hex, ".");
2804 strcpy(temp->mode, ".");
2805 return temp;
2806 }
2807
2808 if (!remove_tempfile_installed) {
2809 atexit(remove_tempfile);
2810 sigchain_push_common(remove_tempfile_on_signal);
2811 remove_tempfile_installed = 1;
2812 }
2813
2814 if (!one->sha1_valid ||
2815 reuse_worktree_file(name, one->sha1, 1)) {
2816 struct stat st;
2817 if (lstat(name, &st) < 0) {
2818 if (errno == ENOENT)
2819 goto not_a_valid_file;
2820 die_errno("stat(%s)", name);
2821 }
2822 if (S_ISLNK(st.st_mode)) {
2823 struct strbuf sb = STRBUF_INIT;
2824 if (strbuf_readlink(&sb, name, st.st_size) < 0)
2825 die_errno("readlink(%s)", name);
2826 prep_temp_blob(name, temp, sb.buf, sb.len,
2827 (one->sha1_valid ?
2828 one->sha1 : null_sha1),
2829 (one->sha1_valid ?
2830 one->mode : S_IFLNK));
2831 strbuf_release(&sb);
2832 }
2833 else {
2834 /* we can borrow from the file in the work tree */
2835 temp->name = name;
2836 if (!one->sha1_valid)
2837 strcpy(temp->hex, sha1_to_hex(null_sha1));
2838 else
2839 strcpy(temp->hex, sha1_to_hex(one->sha1));
2840 /* Even though we may sometimes borrow the
2841 * contents from the work tree, we always want
2842 * one->mode. mode is trustworthy even when
2843 * !(one->sha1_valid), as long as
2844 * DIFF_FILE_VALID(one).
2845 */
2846 sprintf(temp->mode, "%06o", one->mode);
2847 }
2848 return temp;
2849 }
2850 else {
2851 if (diff_populate_filespec(one, 0))
2852 die("cannot read data blob for %s", one->path);
2853 prep_temp_blob(name, temp, one->data, one->size,
2854 one->sha1, one->mode);
2855 }
2856 return temp;
2857 }
2858
2859 /* An external diff command takes:
2860 *
2861 * diff-cmd name infile1 infile1-sha1 infile1-mode \
2862 * infile2 infile2-sha1 infile2-mode [ rename-to ]
2863 *
2864 */
2865 static void run_external_diff(const char *pgm,
2866 const char *name,
2867 const char *other,
2868 struct diff_filespec *one,
2869 struct diff_filespec *two,
2870 const char *xfrm_msg,
2871 int complete_rewrite)
2872 {
2873 const char *spawn_arg[10];
2874 int retval;
2875 const char **arg = &spawn_arg[0];
2876
2877 if (one && two) {
2878 struct diff_tempfile *temp_one, *temp_two;
2879 const char *othername = (other ? other : name);
2880 temp_one = prepare_temp_file(name, one);
2881 temp_two = prepare_temp_file(othername, two);
2882 *arg++ = pgm;
2883 *arg++ = name;
2884 *arg++ = temp_one->name;
2885 *arg++ = temp_one->hex;
2886 *arg++ = temp_one->mode;
2887 *arg++ = temp_two->name;
2888 *arg++ = temp_two->hex;
2889 *arg++ = temp_two->mode;
2890 if (other) {
2891 *arg++ = other;
2892 *arg++ = xfrm_msg;
2893 }
2894 } else {
2895 *arg++ = pgm;
2896 *arg++ = name;
2897 }
2898 *arg = NULL;
2899 fflush(NULL);
2900 retval = run_command_v_opt(spawn_arg, RUN_USING_SHELL);
2901 remove_tempfile();
2902 if (retval) {
2903 fprintf(stderr, "external diff died, stopping at %s.\n", name);
2904 exit(1);
2905 }
2906 }
2907
2908 static int similarity_index(struct diff_filepair *p)
2909 {
2910 return p->score * 100 / MAX_SCORE;
2911 }
2912
2913 static void fill_metainfo(struct strbuf *msg,
2914 const char *name,
2915 const char *other,
2916 struct diff_filespec *one,
2917 struct diff_filespec *two,
2918 struct diff_options *o,
2919 struct diff_filepair *p,
2920 int *must_show_header,
2921 int use_color)
2922 {
2923 const char *set = diff_get_color(use_color, DIFF_METAINFO);
2924 const char *reset = diff_get_color(use_color, DIFF_RESET);
2925 struct strbuf *msgbuf;
2926 char *line_prefix = "";
2927
2928 *must_show_header = 1;
2929 if (o->output_prefix) {
2930 msgbuf = o->output_prefix(o, o->output_prefix_data);
2931 line_prefix = msgbuf->buf;
2932 }
2933 strbuf_init(msg, PATH_MAX * 2 + 300);
2934 switch (p->status) {
2935 case DIFF_STATUS_COPIED:
2936 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2937 line_prefix, set, similarity_index(p));
2938 strbuf_addf(msg, "%s\n%s%scopy from ",
2939 reset, line_prefix, set);
2940 quote_c_style(name, msg, NULL, 0);
2941 strbuf_addf(msg, "%s\n%s%scopy to ", reset, line_prefix, set);
2942 quote_c_style(other, msg, NULL, 0);
2943 strbuf_addf(msg, "%s\n", reset);
2944 break;
2945 case DIFF_STATUS_RENAMED:
2946 strbuf_addf(msg, "%s%ssimilarity index %d%%",
2947 line_prefix, set, similarity_index(p));
2948 strbuf_addf(msg, "%s\n%s%srename from ",
2949 reset, line_prefix, set);
2950 quote_c_style(name, msg, NULL, 0);
2951 strbuf_addf(msg, "%s\n%s%srename to ",
2952 reset, line_prefix, set);
2953 quote_c_style(other, msg, NULL, 0);
2954 strbuf_addf(msg, "%s\n", reset);
2955 break;
2956 case DIFF_STATUS_MODIFIED:
2957 if (p->score) {
2958 strbuf_addf(msg, "%s%sdissimilarity index %d%%%s\n",
2959 line_prefix,
2960 set, similarity_index(p), reset);
2961 break;
2962 }
2963 /* fallthru */
2964 default:
2965 *must_show_header = 0;
2966 }
2967 if (one && two && hashcmp(one->sha1, two->sha1)) {
2968 int abbrev = DIFF_OPT_TST(o, FULL_INDEX) ? 40 : DEFAULT_ABBREV;
2969
2970 if (DIFF_OPT_TST(o, BINARY)) {
2971 mmfile_t mf;
2972 if ((!fill_mmfile(&mf, one) && diff_filespec_is_binary(one)) ||
2973 (!fill_mmfile(&mf, two) && diff_filespec_is_binary(two)))
2974 abbrev = 40;
2975 }
2976 strbuf_addf(msg, "%s%sindex %s..", line_prefix, set,
2977 find_unique_abbrev(one->sha1, abbrev));
2978 strbuf_addstr(msg, find_unique_abbrev(two->sha1, abbrev));
2979 if (one->mode == two->mode)
2980 strbuf_addf(msg, " %06o", one->mode);
2981 strbuf_addf(msg, "%s\n", reset);
2982 }
2983 }
2984
2985 static void run_diff_cmd(const char *pgm,
2986 const char *name,
2987 const char *other,
2988 const char *attr_path,
2989 struct diff_filespec *one,
2990 struct diff_filespec *two,
2991 struct strbuf *msg,
2992 struct diff_options *o,
2993 struct diff_filepair *p)
2994 {
2995 const char *xfrm_msg = NULL;
2996 int complete_rewrite = (p->status == DIFF_STATUS_MODIFIED) && p->score;
2997 int must_show_header = 0;
2998
2999 if (!DIFF_OPT_TST(o, ALLOW_EXTERNAL))
3000 pgm = NULL;
3001 else {
3002 struct userdiff_driver *drv = userdiff_find_by_path(attr_path);
3003 if (drv && drv->external)
3004 pgm = drv->external;
3005 }
3006
3007 if (msg) {
3008 /*
3009 * don't use colors when the header is intended for an
3010 * external diff driver
3011 */
3012 fill_metainfo(msg, name, other, one, two, o, p,
3013 &must_show_header,
3014 want_color(o->use_color) && !pgm);
3015 xfrm_msg = msg->len ? msg->buf : NULL;
3016 }
3017
3018 if (pgm) {
3019 run_external_diff(pgm, name, other, one, two, xfrm_msg,
3020 complete_rewrite);
3021 return;
3022 }
3023 if (one && two)
3024 builtin_diff(name, other ? other : name,
3025 one, two, xfrm_msg, must_show_header,
3026 o, complete_rewrite);
3027 else
3028 fprintf(o->file, "* Unmerged path %s\n", name);
3029 }
3030
3031 static void diff_fill_sha1_info(struct diff_filespec *one)
3032 {
3033 if (DIFF_FILE_VALID(one)) {
3034 if (!one->sha1_valid) {
3035 struct stat st;
3036 if (!strcmp(one->path, "-")) {
3037 hashcpy(one->sha1, null_sha1);
3038 return;
3039 }
3040 if (lstat(one->path, &st) < 0)
3041 die_errno("stat '%s'", one->path);
3042 if (index_path(one->sha1, one->path, &st, 0))
3043 die("cannot hash %s", one->path);
3044 }
3045 }
3046 else
3047 hashclr(one->sha1);
3048 }
3049
3050 static void strip_prefix(int prefix_length, const char **namep, const char **otherp)
3051 {
3052 /* Strip the prefix but do not molest /dev/null and absolute paths */
3053 if (*namep && **namep != '/') {
3054 *namep += prefix_length;
3055 if (**namep == '/')
3056 ++*namep;
3057 }
3058 if (*otherp && **otherp != '/') {
3059 *otherp += prefix_length;
3060 if (**otherp == '/')
3061 ++*otherp;
3062 }
3063 }
3064
3065 static void run_diff(struct diff_filepair *p, struct diff_options *o)
3066 {
3067 const char *pgm = external_diff();
3068 struct strbuf msg;
3069 struct diff_filespec *one = p->one;
3070 struct diff_filespec *two = p->two;
3071 const char *name;
3072 const char *other;
3073 const char *attr_path;
3074
3075 name = p->one->path;
3076 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3077 attr_path = name;
3078 if (o->prefix_length)
3079 strip_prefix(o->prefix_length, &name, &other);
3080
3081 if (DIFF_PAIR_UNMERGED(p)) {
3082 run_diff_cmd(pgm, name, NULL, attr_path,
3083 NULL, NULL, NULL, o, p);
3084 return;
3085 }
3086
3087 diff_fill_sha1_info(one);
3088 diff_fill_sha1_info(two);
3089
3090 if (!pgm &&
3091 DIFF_FILE_VALID(one) && DIFF_FILE_VALID(two) &&
3092 (S_IFMT & one->mode) != (S_IFMT & two->mode)) {
3093 /*
3094 * a filepair that changes between file and symlink
3095 * needs to be split into deletion and creation.
3096 */
3097 struct diff_filespec *null = alloc_filespec(two->path);
3098 run_diff_cmd(NULL, name, other, attr_path,
3099 one, null, &msg, o, p);
3100 free(null);
3101 strbuf_release(&msg);
3102
3103 null = alloc_filespec(one->path);
3104 run_diff_cmd(NULL, name, other, attr_path,
3105 null, two, &msg, o, p);
3106 free(null);
3107 }
3108 else
3109 run_diff_cmd(pgm, name, other, attr_path,
3110 one, two, &msg, o, p);
3111
3112 strbuf_release(&msg);
3113 }
3114
3115 static void run_diffstat(struct diff_filepair *p, struct diff_options *o,
3116 struct diffstat_t *diffstat)
3117 {
3118 const char *name;
3119 const char *other;
3120 int complete_rewrite = 0;
3121
3122 if (DIFF_PAIR_UNMERGED(p)) {
3123 /* unmerged */
3124 builtin_diffstat(p->one->path, NULL, NULL, NULL, diffstat, o, 0);
3125 return;
3126 }
3127
3128 name = p->one->path;
3129 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3130
3131 if (o->prefix_length)
3132 strip_prefix(o->prefix_length, &name, &other);
3133
3134 diff_fill_sha1_info(p->one);
3135 diff_fill_sha1_info(p->two);
3136
3137 if (p->status == DIFF_STATUS_MODIFIED && p->score)
3138 complete_rewrite = 1;
3139 builtin_diffstat(name, other, p->one, p->two, diffstat, o, complete_rewrite);
3140 }
3141
3142 static void run_checkdiff(struct diff_filepair *p, struct diff_options *o)
3143 {
3144 const char *name;
3145 const char *other;
3146 const char *attr_path;
3147
3148 if (DIFF_PAIR_UNMERGED(p)) {
3149 /* unmerged */
3150 return;
3151 }
3152
3153 name = p->one->path;
3154 other = (strcmp(name, p->two->path) ? p->two->path : NULL);
3155 attr_path = other ? other : name;
3156
3157 if (o->prefix_length)
3158 strip_prefix(o->prefix_length, &name, &other);
3159
3160 diff_fill_sha1_info(p->one);
3161 diff_fill_sha1_info(p->two);
3162
3163 builtin_checkdiff(name, other, attr_path, p->one, p->two, o);
3164 }
3165
3166 void diff_setup(struct diff_options *options)
3167 {
3168 memcpy(options, &default_diff_options, sizeof(*options));
3169
3170 options->file = stdout;
3171
3172 options->line_termination = '\n';
3173 options->break_opt = -1;
3174 options->rename_limit = -1;
3175 options->dirstat_permille = diff_dirstat_permille_default;
3176 options->context = 3;
3177 DIFF_OPT_SET(options, RENAME_EMPTY);
3178
3179 options->change = diff_change;
3180 options->add_remove = diff_addremove;
3181 options->use_color = diff_use_color_default;
3182 options->detect_rename = diff_detect_rename_default;
3183
3184 if (diff_no_prefix) {
3185 options->a_prefix = options->b_prefix = "";
3186 } else if (!diff_mnemonic_prefix) {
3187 options->a_prefix = "a/";
3188 options->b_prefix = "b/";
3189 }
3190 }
3191
3192 int diff_setup_done(struct diff_options *options)
3193 {
3194 int count = 0;
3195
3196 if (options->output_format & DIFF_FORMAT_NAME)
3197 count++;
3198 if (options->output_format & DIFF_FORMAT_NAME_STATUS)
3199 count++;
3200 if (options->output_format & DIFF_FORMAT_CHECKDIFF)
3201 count++;
3202 if (options->output_format & DIFF_FORMAT_NO_OUTPUT)
3203 count++;
3204 if (count > 1)
3205 die("--name-only, --name-status, --check and -s are mutually exclusive");
3206
3207 /*
3208 * Most of the time we can say "there are changes"
3209 * only by checking if there are changed paths, but
3210 * --ignore-whitespace* options force us to look
3211 * inside contents.
3212 */
3213
3214 if (DIFF_XDL_TST(options, IGNORE_WHITESPACE) ||
3215 DIFF_XDL_TST(options, IGNORE_WHITESPACE_CHANGE) ||
3216 DIFF_XDL_TST(options, IGNORE_WHITESPACE_AT_EOL))
3217 DIFF_OPT_SET(options, DIFF_FROM_CONTENTS);
3218 else
3219 DIFF_OPT_CLR(options, DIFF_FROM_CONTENTS);
3220
3221 if (DIFF_OPT_TST(options, FIND_COPIES_HARDER))
3222 options->detect_rename = DIFF_DETECT_COPY;
3223
3224 if (!DIFF_OPT_TST(options, RELATIVE_NAME))
3225 options->prefix = NULL;
3226 if (options->prefix)
3227 options->prefix_length = strlen(options->prefix);
3228 else
3229 options->prefix_length = 0;
3230
3231 if (options->output_format & (DIFF_FORMAT_NAME |
3232 DIFF_FORMAT_NAME_STATUS |
3233 DIFF_FORMAT_CHECKDIFF |
3234 DIFF_FORMAT_NO_OUTPUT))
3235 options->output_format &= ~(DIFF_FORMAT_RAW |
3236 DIFF_FORMAT_NUMSTAT |
3237 DIFF_FORMAT_DIFFSTAT |
3238 DIFF_FORMAT_SHORTSTAT |
3239 DIFF_FORMAT_DIRSTAT |
3240 DIFF_FORMAT_SUMMARY |
3241 DIFF_FORMAT_PATCH);
3242
3243 /*
3244 * These cases always need recursive; we do not drop caller-supplied
3245 * recursive bits for other formats here.
3246 */
3247 if (options->output_format & (DIFF_FORMAT_PATCH |
3248 DIFF_FORMAT_NUMSTAT |
3249 DIFF_FORMAT_DIFFSTAT |
3250 DIFF_FORMAT_SHORTSTAT |
3251 DIFF_FORMAT_DIRSTAT |
3252 DIFF_FORMAT_SUMMARY |
3253 DIFF_FORMAT_CHECKDIFF))
3254 DIFF_OPT_SET(options, RECURSIVE);
3255 /*
3256 * Also pickaxe would not work very well if you do not say recursive
3257 */
3258 if (options->pickaxe)
3259 DIFF_OPT_SET(options, RECURSIVE);
3260 /*
3261 * When patches are generated, submodules diffed against the work tree
3262 * must be checked for dirtiness too so it can be shown in the output
3263 */
3264 if (options->output_format & DIFF_FORMAT_PATCH)
3265 DIFF_OPT_SET(options, DIRTY_SUBMODULES);
3266
3267 if (options->detect_rename && options->rename_limit < 0)
3268 options->rename_limit = diff_rename_limit_default;
3269 if (options->setup & DIFF_SETUP_USE_CACHE) {
3270 if (!active_cache)
3271 /* read-cache does not die even when it fails
3272 * so it is safe for us to do this here. Also
3273 * it does not smudge active_cache or active_nr
3274 * when it fails, so we do not have to worry about
3275 * cleaning it up ourselves either.
3276 */
3277 read_cache();
3278 }
3279 if (options->abbrev <= 0 || 40 < options->abbrev)
3280 options->abbrev = 40; /* full */
3281
3282 /*
3283 * It does not make sense to show the first hit we happened
3284 * to have found. It does not make sense not to return with
3285 * exit code in such a case either.
3286 */
3287 if (DIFF_OPT_TST(options, QUICK)) {
3288 options->output_format = DIFF_FORMAT_NO_OUTPUT;
3289 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3290 }
3291
3292 return 0;
3293 }
3294
3295 static int opt_arg(const char *arg, int arg_short, const char *arg_long, int *val)
3296 {
3297 char c, *eq;
3298 int len;
3299
3300 if (*arg != '-')
3301 return 0;
3302 c = *++arg;
3303 if (!c)
3304 return 0;
3305 if (c == arg_short) {
3306 c = *++arg;
3307 if (!c)
3308 return 1;
3309 if (val && isdigit(c)) {
3310 char *end;
3311 int n = strtoul(arg, &end, 10);
3312 if (*end)
3313 return 0;
3314 *val = n;
3315 return 1;
3316 }
3317 return 0;
3318 }
3319 if (c != '-')
3320 return 0;
3321 arg++;
3322 eq = strchr(arg, '=');
3323 if (eq)
3324 len = eq - arg;
3325 else
3326 len = strlen(arg);
3327 if (!len || strncmp(arg, arg_long, len))
3328 return 0;
3329 if (eq) {
3330 int n;
3331 char *end;
3332 if (!isdigit(*++eq))
3333 return 0;
3334 n = strtoul(eq, &end, 10);
3335 if (*end)
3336 return 0;
3337 *val = n;
3338 }
3339 return 1;
3340 }
3341
3342 static int diff_scoreopt_parse(const char *opt);
3343
3344 static inline int short_opt(char opt, const char **argv,
3345 const char **optarg)
3346 {
3347 const char *arg = argv[0];
3348 if (arg[0] != '-' || arg[1] != opt)
3349 return 0;
3350 if (arg[2] != '\0') {
3351 *optarg = arg + 2;
3352 return 1;
3353 }
3354 if (!argv[1])
3355 die("Option '%c' requires a value", opt);
3356 *optarg = argv[1];
3357 return 2;
3358 }
3359
3360 int parse_long_opt(const char *opt, const char **argv,
3361 const char **optarg)
3362 {
3363 const char *arg = argv[0];
3364 if (arg[0] != '-' || arg[1] != '-')
3365 return 0;
3366 arg += strlen("--");
3367 if (prefixcmp(arg, opt))
3368 return 0;
3369 arg += strlen(opt);
3370 if (*arg == '=') { /* sticked form: --option=value */
3371 *optarg = arg + 1;
3372 return 1;
3373 }
3374 if (*arg != '\0')
3375 return 0;
3376 /* separate form: --option value */
3377 if (!argv[1])
3378 die("Option '--%s' requires a value", opt);
3379 *optarg = argv[1];
3380 return 2;
3381 }
3382
3383 static int stat_opt(struct diff_options *options, const char **av)
3384 {
3385 const char *arg = av[0];
3386 char *end;
3387 int width = options->stat_width;
3388 int name_width = options->stat_name_width;
3389 int graph_width = options->stat_graph_width;
3390 int count = options->stat_count;
3391 int argcount = 1;
3392
3393 arg += strlen("--stat");
3394 end = (char *)arg;
3395
3396 switch (*arg) {
3397 case '-':
3398 if (!prefixcmp(arg, "-width")) {
3399 arg += strlen("-width");
3400 if (*arg == '=')
3401 width = strtoul(arg + 1, &end, 10);
3402 else if (!*arg && !av[1])
3403 die("Option '--stat-width' requires a value");
3404 else if (!*arg) {
3405 width = strtoul(av[1], &end, 10);
3406 argcount = 2;
3407 }
3408 } else if (!prefixcmp(arg, "-name-width")) {
3409 arg += strlen("-name-width");
3410 if (*arg == '=')
3411 name_width = strtoul(arg + 1, &end, 10);
3412 else if (!*arg && !av[1])
3413 die("Option '--stat-name-width' requires a value");
3414 else if (!*arg) {
3415 name_width = strtoul(av[1], &end, 10);
3416 argcount = 2;
3417 }
3418 } else if (!prefixcmp(arg, "-graph-width")) {
3419 arg += strlen("-graph-width");
3420 if (*arg == '=')
3421 graph_width = strtoul(arg + 1, &end, 10);
3422 else if (!*arg && !av[1])
3423 die("Option '--stat-graph-width' requires a value");
3424 else if (!*arg) {
3425 graph_width = strtoul(av[1], &end, 10);
3426 argcount = 2;
3427 }
3428 } else if (!prefixcmp(arg, "-count")) {
3429 arg += strlen("-count");
3430 if (*arg == '=')
3431 count = strtoul(arg + 1, &end, 10);
3432 else if (!*arg && !av[1])
3433 die("Option '--stat-count' requires a value");
3434 else if (!*arg) {
3435 count = strtoul(av[1], &end, 10);
3436 argcount = 2;
3437 }
3438 }
3439 break;
3440 case '=':
3441 width = strtoul(arg+1, &end, 10);
3442 if (*end == ',')
3443 name_width = strtoul(end+1, &end, 10);
3444 if (*end == ',')
3445 count = strtoul(end+1, &end, 10);
3446 }
3447
3448 /* Important! This checks all the error cases! */
3449 if (*end)
3450 return 0;
3451 options->output_format |= DIFF_FORMAT_DIFFSTAT;
3452 options->stat_name_width = name_width;
3453 options->stat_graph_width = graph_width;
3454 options->stat_width = width;
3455 options->stat_count = count;
3456 return argcount;
3457 }
3458
3459 static int parse_dirstat_opt(struct diff_options *options, const char *params)
3460 {
3461 struct strbuf errmsg = STRBUF_INIT;
3462 if (parse_dirstat_params(options, params, &errmsg))
3463 die(_("Failed to parse --dirstat/-X option parameter:\n%s"),
3464 errmsg.buf);
3465 strbuf_release(&errmsg);
3466 /*
3467 * The caller knows a dirstat-related option is given from the command
3468 * line; allow it to say "return this_function();"
3469 */
3470 options->output_format |= DIFF_FORMAT_DIRSTAT;
3471 return 1;
3472 }
3473
3474 int diff_opt_parse(struct diff_options *options, const char **av, int ac)
3475 {
3476 const char *arg = av[0];
3477 const char *optarg;
3478 int argcount;
3479
3480 /* Output format options */
3481 if (!strcmp(arg, "-p") || !strcmp(arg, "-u") || !strcmp(arg, "--patch"))
3482 options->output_format |= DIFF_FORMAT_PATCH;
3483 else if (opt_arg(arg, 'U', "unified", &options->context))
3484 options->output_format |= DIFF_FORMAT_PATCH;
3485 else if (!strcmp(arg, "--raw"))
3486 options->output_format |= DIFF_FORMAT_RAW;
3487 else if (!strcmp(arg, "--patch-with-raw"))
3488 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_RAW;
3489 else if (!strcmp(arg, "--numstat"))
3490 options->output_format |= DIFF_FORMAT_NUMSTAT;
3491 else if (!strcmp(arg, "--shortstat"))
3492 options->output_format |= DIFF_FORMAT_SHORTSTAT;
3493 else if (!strcmp(arg, "-X") || !strcmp(arg, "--dirstat"))
3494 return parse_dirstat_opt(options, "");
3495 else if (!prefixcmp(arg, "-X"))
3496 return parse_dirstat_opt(options, arg + 2);
3497 else if (!prefixcmp(arg, "--dirstat="))
3498 return parse_dirstat_opt(options, arg + 10);
3499 else if (!strcmp(arg, "--cumulative"))
3500 return parse_dirstat_opt(options, "cumulative");
3501 else if (!strcmp(arg, "--dirstat-by-file"))
3502 return parse_dirstat_opt(options, "files");
3503 else if (!prefixcmp(arg, "--dirstat-by-file=")) {
3504 parse_dirstat_opt(options, "files");
3505 return parse_dirstat_opt(options, arg + 18);
3506 }
3507 else if (!strcmp(arg, "--check"))
3508 options->output_format |= DIFF_FORMAT_CHECKDIFF;
3509 else if (!strcmp(arg, "--summary"))
3510 options->output_format |= DIFF_FORMAT_SUMMARY;
3511 else if (!strcmp(arg, "--patch-with-stat"))
3512 options->output_format |= DIFF_FORMAT_PATCH | DIFF_FORMAT_DIFFSTAT;
3513 else if (!strcmp(arg, "--name-only"))
3514 options->output_format |= DIFF_FORMAT_NAME;
3515 else if (!strcmp(arg, "--name-status"))
3516 options->output_format |= DIFF_FORMAT_NAME_STATUS;
3517 else if (!strcmp(arg, "-s"))
3518 options->output_format |= DIFF_FORMAT_NO_OUTPUT;
3519 else if (!prefixcmp(arg, "--stat"))
3520 /* --stat, --stat-width, --stat-name-width, or --stat-count */
3521 return stat_opt(options, av);
3522
3523 /* renames options */
3524 else if (!prefixcmp(arg, "-B") || !prefixcmp(arg, "--break-rewrites=") ||
3525 !strcmp(arg, "--break-rewrites")) {
3526 if ((options->break_opt = diff_scoreopt_parse(arg)) == -1)
3527 return error("invalid argument to -B: %s", arg+2);
3528 }
3529 else if (!prefixcmp(arg, "-M") || !prefixcmp(arg, "--find-renames=") ||
3530 !strcmp(arg, "--find-renames")) {
3531 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3532 return error("invalid argument to -M: %s", arg+2);
3533 options->detect_rename = DIFF_DETECT_RENAME;
3534 }
3535 else if (!strcmp(arg, "-D") || !strcmp(arg, "--irreversible-delete")) {
3536 options->irreversible_delete = 1;
3537 }
3538 else if (!prefixcmp(arg, "-C") || !prefixcmp(arg, "--find-copies=") ||
3539 !strcmp(arg, "--find-copies")) {
3540 if (options->detect_rename == DIFF_DETECT_COPY)
3541 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3542 if ((options->rename_score = diff_scoreopt_parse(arg)) == -1)
3543 return error("invalid argument to -C: %s", arg+2);
3544 options->detect_rename = DIFF_DETECT_COPY;
3545 }
3546 else if (!strcmp(arg, "--no-renames"))
3547 options->detect_rename = 0;
3548 else if (!strcmp(arg, "--rename-empty"))
3549 DIFF_OPT_SET(options, RENAME_EMPTY);
3550 else if (!strcmp(arg, "--no-rename-empty"))
3551 DIFF_OPT_CLR(options, RENAME_EMPTY);
3552 else if (!strcmp(arg, "--relative"))
3553 DIFF_OPT_SET(options, RELATIVE_NAME);
3554 else if (!prefixcmp(arg, "--relative=")) {
3555 DIFF_OPT_SET(options, RELATIVE_NAME);
3556 options->prefix = arg + 11;
3557 }
3558
3559 /* xdiff options */
3560 else if (!strcmp(arg, "--minimal"))
3561 DIFF_XDL_SET(options, NEED_MINIMAL);
3562 else if (!strcmp(arg, "--no-minimal"))
3563 DIFF_XDL_CLR(options, NEED_MINIMAL);
3564 else if (!strcmp(arg, "-w") || !strcmp(arg, "--ignore-all-space"))
3565 DIFF_XDL_SET(options, IGNORE_WHITESPACE);
3566 else if (!strcmp(arg, "-b") || !strcmp(arg, "--ignore-space-change"))
3567 DIFF_XDL_SET(options, IGNORE_WHITESPACE_CHANGE);
3568 else if (!strcmp(arg, "--ignore-space-at-eol"))
3569 DIFF_XDL_SET(options, IGNORE_WHITESPACE_AT_EOL);
3570 else if (!strcmp(arg, "--patience"))
3571 options->xdl_opts = DIFF_WITH_ALG(options, PATIENCE_DIFF);
3572 else if (!strcmp(arg, "--histogram"))
3573 options->xdl_opts = DIFF_WITH_ALG(options, HISTOGRAM_DIFF);
3574
3575 /* flags options */
3576 else if (!strcmp(arg, "--binary")) {
3577 options->output_format |= DIFF_FORMAT_PATCH;
3578 DIFF_OPT_SET(options, BINARY);
3579 }
3580 else if (!strcmp(arg, "--full-index"))
3581 DIFF_OPT_SET(options, FULL_INDEX);
3582 else if (!strcmp(arg, "-a") || !strcmp(arg, "--text"))
3583 DIFF_OPT_SET(options, TEXT);
3584 else if (!strcmp(arg, "-R"))
3585 DIFF_OPT_SET(options, REVERSE_DIFF);
3586 else if (!strcmp(arg, "--find-copies-harder"))
3587 DIFF_OPT_SET(options, FIND_COPIES_HARDER);
3588 else if (!strcmp(arg, "--follow"))
3589 DIFF_OPT_SET(options, FOLLOW_RENAMES);
3590 else if (!strcmp(arg, "--color"))
3591 options->use_color = 1;
3592 else if (!prefixcmp(arg, "--color=")) {
3593 int value = git_config_colorbool(NULL, arg+8);
3594 if (value < 0)
3595 return error("option `color' expects \"always\", \"auto\", or \"never\"");
3596 options->use_color = value;
3597 }
3598 else if (!strcmp(arg, "--no-color"))
3599 options->use_color = 0;
3600 else if (!strcmp(arg, "--color-words")) {
3601 options->use_color = 1;
3602 options->word_diff = DIFF_WORDS_COLOR;
3603 }
3604 else if (!prefixcmp(arg, "--color-words=")) {
3605 options->use_color = 1;
3606 options->word_diff = DIFF_WORDS_COLOR;
3607 options->word_regex = arg + 14;
3608 }
3609 else if (!strcmp(arg, "--word-diff")) {
3610 if (options->word_diff == DIFF_WORDS_NONE)
3611 options->word_diff = DIFF_WORDS_PLAIN;
3612 }
3613 else if (!prefixcmp(arg, "--word-diff=")) {
3614 const char *type = arg + 12;
3615 if (!strcmp(type, "plain"))
3616 options->word_diff = DIFF_WORDS_PLAIN;
3617 else if (!strcmp(type, "color")) {
3618 options->use_color = 1;
3619 options->word_diff = DIFF_WORDS_COLOR;
3620 }
3621 else if (!strcmp(type, "porcelain"))
3622 options->word_diff = DIFF_WORDS_PORCELAIN;
3623 else if (!strcmp(type, "none"))
3624 options->word_diff = DIFF_WORDS_NONE;
3625 else
3626 die("bad --word-diff argument: %s", type);
3627 }
3628 else if ((argcount = parse_long_opt("word-diff-regex", av, &optarg))) {
3629 if (options->word_diff == DIFF_WORDS_NONE)
3630 options->word_diff = DIFF_WORDS_PLAIN;
3631 options->word_regex = optarg;
3632 return argcount;
3633 }
3634 else if (!strcmp(arg, "--exit-code"))
3635 DIFF_OPT_SET(options, EXIT_WITH_STATUS);
3636 else if (!strcmp(arg, "--quiet"))
3637 DIFF_OPT_SET(options, QUICK);
3638 else if (!strcmp(arg, "--ext-diff"))
3639 DIFF_OPT_SET(options, ALLOW_EXTERNAL);
3640 else if (!strcmp(arg, "--no-ext-diff"))
3641 DIFF_OPT_CLR(options, ALLOW_EXTERNAL);
3642 else if (!strcmp(arg, "--textconv"))
3643 DIFF_OPT_SET(options, ALLOW_TEXTCONV);
3644 else if (!strcmp(arg, "--no-textconv"))
3645 DIFF_OPT_CLR(options, ALLOW_TEXTCONV);
3646 else if (!strcmp(arg, "--ignore-submodules")) {
3647 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3648 handle_ignore_submodules_arg(options, "all");
3649 } else if (!prefixcmp(arg, "--ignore-submodules=")) {
3650 DIFF_OPT_SET(options, OVERRIDE_SUBMODULE_CONFIG);
3651 handle_ignore_submodules_arg(options, arg + 20);
3652 } else if (!strcmp(arg, "--submodule"))
3653 DIFF_OPT_SET(options, SUBMODULE_LOG);
3654 else if (!prefixcmp(arg, "--submodule=")) {
3655 if (!strcmp(arg + 12, "log"))
3656 DIFF_OPT_SET(options, SUBMODULE_LOG);
3657 }
3658
3659 /* misc options */
3660 else if (!strcmp(arg, "-z"))
3661 options->line_termination = 0;
3662 else if ((argcount = short_opt('l', av, &optarg))) {
3663 options->rename_limit = strtoul(optarg, NULL, 10);
3664 return argcount;
3665 }
3666 else if ((argcount = short_opt('S', av, &optarg))) {
3667 options->pickaxe = optarg;
3668 options->pickaxe_opts |= DIFF_PICKAXE_KIND_S;
3669 return argcount;
3670 } else if ((argcount = short_opt('G', av, &optarg))) {
3671 options->pickaxe = optarg;
3672 options->pickaxe_opts |= DIFF_PICKAXE_KIND_G;
3673 return argcount;
3674 }
3675 else if (!strcmp(arg, "--pickaxe-all"))
3676 options->pickaxe_opts |= DIFF_PICKAXE_ALL;
3677 else if (!strcmp(arg, "--pickaxe-regex"))
3678 options->pickaxe_opts |= DIFF_PICKAXE_REGEX;
3679 else if ((argcount = short_opt('O', av, &optarg))) {
3680 options->orderfile = optarg;
3681 return argcount;
3682 }
3683 else if ((argcount = parse_long_opt("diff-filter", av, &optarg))) {
3684 options->filter = optarg;
3685 return argcount;
3686 }
3687 else if (!strcmp(arg, "--abbrev"))
3688 options->abbrev = DEFAULT_ABBREV;
3689 else if (!prefixcmp(arg, "--abbrev=")) {
3690 options->abbrev = strtoul(arg + 9, NULL, 10);
3691 if (options->abbrev < MINIMUM_ABBREV)
3692 options->abbrev = MINIMUM_ABBREV;
3693 else if (40 < options->abbrev)
3694 options->abbrev = 40;
3695 }
3696 else if ((argcount = parse_long_opt("src-prefix", av, &optarg))) {
3697 options->a_prefix = optarg;
3698 return argcount;
3699 }
3700 else if ((argcount = parse_long_opt("dst-prefix", av, &optarg))) {
3701 options->b_prefix = optarg;
3702 return argcount;
3703 }
3704 else if (!strcmp(arg, "--no-prefix"))
3705 options->a_prefix = options->b_prefix = "";
3706 else if (opt_arg(arg, '\0', "inter-hunk-context",
3707 &options->interhunkcontext))
3708 ;
3709 else if (!strcmp(arg, "-W"))
3710 DIFF_OPT_SET(options, FUNCCONTEXT);
3711 else if (!strcmp(arg, "--function-context"))
3712 DIFF_OPT_SET(options, FUNCCONTEXT);
3713 else if (!strcmp(arg, "--no-function-context"))
3714 DIFF_OPT_CLR(options, FUNCCONTEXT);
3715 else if ((argcount = parse_long_opt("output", av, &optarg))) {
3716 options->file = fopen(optarg, "w");
3717 if (!options->file)
3718 die_errno("Could not open '%s'", optarg);
3719 options->close_file = 1;
3720 return argcount;
3721 } else
3722 return 0;
3723 return 1;
3724 }
3725
3726 int parse_rename_score(const char **cp_p)
3727 {
3728 unsigned long num, scale;
3729 int ch, dot;
3730 const char *cp = *cp_p;
3731
3732 num = 0;
3733 scale = 1;
3734 dot = 0;
3735 for (;;) {
3736 ch = *cp;
3737 if ( !dot && ch == '.' ) {
3738 scale = 1;
3739 dot = 1;
3740 } else if ( ch == '%' ) {
3741 scale = dot ? scale*100 : 100;
3742 cp++; /* % is always at the end */
3743 break;
3744 } else if ( ch >= '0' && ch <= '9' ) {
3745 if ( scale < 100000 ) {
3746 scale *= 10;
3747 num = (num*10) + (ch-'0');
3748 }
3749 } else {
3750 break;
3751 }
3752 cp++;
3753 }
3754 *cp_p = cp;
3755
3756 /* user says num divided by scale and we say internally that
3757 * is MAX_SCORE * num / scale.
3758 */
3759 return (int)((num >= scale) ? MAX_SCORE : (MAX_SCORE * num / scale));
3760 }
3761
3762 static int diff_scoreopt_parse(const char *opt)
3763 {
3764 int opt1, opt2, cmd;
3765
3766 if (*opt++ != '-')
3767 return -1;
3768 cmd = *opt++;
3769 if (cmd == '-') {
3770 /* convert the long-form arguments into short-form versions */
3771 if (!prefixcmp(opt, "break-rewrites")) {
3772 opt += strlen("break-rewrites");
3773 if (*opt == 0 || *opt++ == '=')
3774 cmd = 'B';
3775 } else if (!prefixcmp(opt, "find-copies")) {
3776 opt += strlen("find-copies");
3777 if (*opt == 0 || *opt++ == '=')
3778 cmd = 'C';
3779 } else if (!prefixcmp(opt, "find-renames")) {
3780 opt += strlen("find-renames");
3781 if (*opt == 0 || *opt++ == '=')
3782 cmd = 'M';
3783 }
3784 }
3785 if (cmd != 'M' && cmd != 'C' && cmd != 'B')
3786 return -1; /* that is not a -M, -C nor -B option */
3787
3788 opt1 = parse_rename_score(&opt);
3789 if (cmd != 'B')
3790 opt2 = 0;
3791 else {
3792 if (*opt == 0)
3793 opt2 = 0;
3794 else if (*opt != '/')
3795 return -1; /* we expect -B80/99 or -B80 */
3796 else {
3797 opt++;
3798 opt2 = parse_rename_score(&opt);
3799 }
3800 }
3801 if (*opt != 0)
3802 return -1;
3803 return opt1 | (opt2 << 16);
3804 }
3805
3806 struct diff_queue_struct diff_queued_diff;
3807
3808 void diff_q(struct diff_queue_struct *queue, struct diff_filepair *dp)
3809 {
3810 if (queue->alloc <= queue->nr) {
3811 queue->alloc = alloc_nr(queue->alloc);
3812 queue->queue = xrealloc(queue->queue,
3813 sizeof(dp) * queue->alloc);
3814 }
3815 queue->queue[queue->nr++] = dp;
3816 }
3817
3818 struct diff_filepair *diff_queue(struct diff_queue_struct *queue,
3819 struct diff_filespec *one,
3820 struct diff_filespec *two)
3821 {
3822 struct diff_filepair *dp = xcalloc(1, sizeof(*dp));
3823 dp->one = one;
3824 dp->two = two;
3825 if (queue)
3826 diff_q(queue, dp);
3827 return dp;
3828 }
3829
3830 void diff_free_filepair(struct diff_filepair *p)
3831 {
3832 free_filespec(p->one);
3833 free_filespec(p->two);
3834 free(p);
3835 }
3836
3837 /* This is different from find_unique_abbrev() in that
3838 * it stuffs the result with dots for alignment.
3839 */
3840 const char *diff_unique_abbrev(const unsigned char *sha1, int len)
3841 {
3842 int abblen;
3843 const char *abbrev;
3844 if (len == 40)
3845 return sha1_to_hex(sha1);
3846
3847 abbrev = find_unique_abbrev(sha1, len);
3848 abblen = strlen(abbrev);
3849 if (abblen < 37) {
3850 static char hex[41];
3851 if (len < abblen && abblen <= len + 2)
3852 sprintf(hex, "%s%.*s", abbrev, len+3-abblen, "..");
3853 else
3854 sprintf(hex, "%s...", abbrev);
3855 return hex;
3856 }
3857 return sha1_to_hex(sha1);
3858 }
3859
3860 static void diff_flush_raw(struct diff_filepair *p, struct diff_options *opt)
3861 {
3862 int line_termination = opt->line_termination;
3863 int inter_name_termination = line_termination ? '\t' : '\0';
3864 if (opt->output_prefix) {
3865 struct strbuf *msg = NULL;
3866 msg = opt->output_prefix(opt, opt->output_prefix_data);
3867 fprintf(opt->file, "%s", msg->buf);
3868 }
3869
3870 if (!(opt->output_format & DIFF_FORMAT_NAME_STATUS)) {
3871 fprintf(opt->file, ":%06o %06o %s ", p->one->mode, p->two->mode,
3872 diff_unique_abbrev(p->one->sha1, opt->abbrev));
3873 fprintf(opt->file, "%s ", diff_unique_abbrev(p->two->sha1, opt->abbrev));
3874 }
3875 if (p->score) {
3876 fprintf(opt->file, "%c%03d%c", p->status, similarity_index(p),
3877 inter_name_termination);
3878 } else {
3879 fprintf(opt->file, "%c%c", p->status, inter_name_termination);
3880 }
3881
3882 if (p->status == DIFF_STATUS_COPIED ||
3883 p->status == DIFF_STATUS_RENAMED) {
3884 const char *name_a, *name_b;
3885 name_a = p->one->path;
3886 name_b = p->two->path;
3887 strip_prefix(opt->prefix_length, &name_a, &name_b);
3888 write_name_quoted(name_a, opt->file, inter_name_termination);
3889 write_name_quoted(name_b, opt->file, line_termination);
3890 } else {
3891 const char *name_a, *name_b;
3892 name_a = p->one->mode ? p->one->path : p->two->path;
3893 name_b = NULL;
3894 strip_prefix(opt->prefix_length, &name_a, &name_b);
3895 write_name_quoted(name_a, opt->file, line_termination);
3896 }
3897 }
3898
3899 int diff_unmodified_pair(struct diff_filepair *p)
3900 {
3901 /* This function is written stricter than necessary to support
3902 * the currently implemented transformers, but the idea is to
3903 * let transformers to produce diff_filepairs any way they want,
3904 * and filter and clean them up here before producing the output.
3905 */
3906 struct diff_filespec *one = p->one, *two = p->two;
3907
3908 if (DIFF_PAIR_UNMERGED(p))
3909 return 0; /* unmerged is interesting */
3910
3911 /* deletion, addition, mode or type change
3912 * and rename are all interesting.
3913 */
3914 if (DIFF_FILE_VALID(one) != DIFF_FILE_VALID(two) ||
3915 DIFF_PAIR_MODE_CHANGED(p) ||
3916 strcmp(one->path, two->path))
3917 return 0;
3918
3919 /* both are valid and point at the same path. that is, we are
3920 * dealing with a change.
3921 */
3922 if (one->sha1_valid && two->sha1_valid &&
3923 !hashcmp(one->sha1, two->sha1) &&
3924 !one->dirty_submodule && !two->dirty_submodule)
3925 return 1; /* no change */
3926 if (!one->sha1_valid && !two->sha1_valid)
3927 return 1; /* both look at the same file on the filesystem. */
3928 return 0;
3929 }
3930
3931 static void diff_flush_patch(struct diff_filepair *p, struct diff_options *o)
3932 {
3933 if (diff_unmodified_pair(p))
3934 return;
3935
3936 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3937 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3938 return; /* no tree diffs in patch format */
3939
3940 run_diff(p, o);
3941 }
3942
3943 static void diff_flush_stat(struct diff_filepair *p, struct diff_options *o,
3944 struct diffstat_t *diffstat)
3945 {
3946 if (diff_unmodified_pair(p))
3947 return;
3948
3949 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3950 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3951 return; /* no useful stat for tree diffs */
3952
3953 run_diffstat(p, o, diffstat);
3954 }
3955
3956 static void diff_flush_checkdiff(struct diff_filepair *p,
3957 struct diff_options *o)
3958 {
3959 if (diff_unmodified_pair(p))
3960 return;
3961
3962 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
3963 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
3964 return; /* nothing to check in tree diffs */
3965
3966 run_checkdiff(p, o);
3967 }
3968
3969 int diff_queue_is_empty(void)
3970 {
3971 struct diff_queue_struct *q = &diff_queued_diff;
3972 int i;
3973 for (i = 0; i < q->nr; i++)
3974 if (!diff_unmodified_pair(q->queue[i]))
3975 return 0;
3976 return 1;
3977 }
3978
3979 #if DIFF_DEBUG
3980 void diff_debug_filespec(struct diff_filespec *s, int x, const char *one)
3981 {
3982 fprintf(stderr, "queue[%d] %s (%s) %s %06o %s\n",
3983 x, one ? one : "",
3984 s->path,
3985 DIFF_FILE_VALID(s) ? "valid" : "invalid",
3986 s->mode,
3987 s->sha1_valid ? sha1_to_hex(s->sha1) : "");
3988 fprintf(stderr, "queue[%d] %s size %lu flags %d\n",
3989 x, one ? one : "",
3990 s->size, s->xfrm_flags);
3991 }
3992
3993 void diff_debug_filepair(const struct diff_filepair *p, int i)
3994 {
3995 diff_debug_filespec(p->one, i, "one");
3996 diff_debug_filespec(p->two, i, "two");
3997 fprintf(stderr, "score %d, status %c rename_used %d broken %d\n",
3998 p->score, p->status ? p->status : '?',
3999 p->one->rename_used, p->broken_pair);
4000 }
4001
4002 void diff_debug_queue(const char *msg, struct diff_queue_struct *q)
4003 {
4004 int i;
4005 if (msg)
4006 fprintf(stderr, "%s\n", msg);
4007 fprintf(stderr, "q->nr = %d\n", q->nr);
4008 for (i = 0; i < q->nr; i++) {
4009 struct diff_filepair *p = q->queue[i];
4010 diff_debug_filepair(p, i);
4011 }
4012 }
4013 #endif
4014
4015 static void diff_resolve_rename_copy(void)
4016 {
4017 int i;
4018 struct diff_filepair *p;
4019 struct diff_queue_struct *q = &diff_queued_diff;
4020
4021 diff_debug_queue("resolve-rename-copy", q);
4022
4023 for (i = 0; i < q->nr; i++) {
4024 p = q->queue[i];
4025 p->status = 0; /* undecided */
4026 if (DIFF_PAIR_UNMERGED(p))
4027 p->status = DIFF_STATUS_UNMERGED;
4028 else if (!DIFF_FILE_VALID(p->one))
4029 p->status = DIFF_STATUS_ADDED;
4030 else if (!DIFF_FILE_VALID(p->two))
4031 p->status = DIFF_STATUS_DELETED;
4032 else if (DIFF_PAIR_TYPE_CHANGED(p))
4033 p->status = DIFF_STATUS_TYPE_CHANGED;
4034
4035 /* from this point on, we are dealing with a pair
4036 * whose both sides are valid and of the same type, i.e.
4037 * either in-place edit or rename/copy edit.
4038 */
4039 else if (DIFF_PAIR_RENAME(p)) {
4040 /*
4041 * A rename might have re-connected a broken
4042 * pair up, causing the pathnames to be the
4043 * same again. If so, that's not a rename at
4044 * all, just a modification..
4045 *
4046 * Otherwise, see if this source was used for
4047 * multiple renames, in which case we decrement
4048 * the count, and call it a copy.
4049 */
4050 if (!strcmp(p->one->path, p->two->path))
4051 p->status = DIFF_STATUS_MODIFIED;
4052 else if (--p->one->rename_used > 0)
4053 p->status = DIFF_STATUS_COPIED;
4054 else
4055 p->status = DIFF_STATUS_RENAMED;
4056 }
4057 else if (hashcmp(p->one->sha1, p->two->sha1) ||
4058 p->one->mode != p->two->mode ||
4059 p->one->dirty_submodule ||
4060 p->two->dirty_submodule ||
4061 is_null_sha1(p->one->sha1))
4062 p->status = DIFF_STATUS_MODIFIED;
4063 else {
4064 /* This is a "no-change" entry and should not
4065 * happen anymore, but prepare for broken callers.
4066 */
4067 error("feeding unmodified %s to diffcore",
4068 p->one->path);
4069 p->status = DIFF_STATUS_UNKNOWN;
4070 }
4071 }
4072 diff_debug_queue("resolve-rename-copy done", q);
4073 }
4074
4075 static int check_pair_status(struct diff_filepair *p)
4076 {
4077 switch (p->status) {
4078 case DIFF_STATUS_UNKNOWN:
4079 return 0;
4080 case 0:
4081 die("internal error in diff-resolve-rename-copy");
4082 default:
4083 return 1;
4084 }
4085 }
4086
4087 static void flush_one_pair(struct diff_filepair *p, struct diff_options *opt)
4088 {
4089 int fmt = opt->output_format;
4090
4091 if (fmt & DIFF_FORMAT_CHECKDIFF)
4092 diff_flush_checkdiff(p, opt);
4093 else if (fmt & (DIFF_FORMAT_RAW | DIFF_FORMAT_NAME_STATUS))
4094 diff_flush_raw(p, opt);
4095 else if (fmt & DIFF_FORMAT_NAME) {
4096 const char *name_a, *name_b;
4097 name_a = p->two->path;
4098 name_b = NULL;
4099 strip_prefix(opt->prefix_length, &name_a, &name_b);
4100 write_name_quoted(name_a, opt->file, opt->line_termination);
4101 }
4102 }
4103
4104 static void show_file_mode_name(FILE *file, const char *newdelete, struct diff_filespec *fs)
4105 {
4106 if (fs->mode)
4107 fprintf(file, " %s mode %06o ", newdelete, fs->mode);
4108 else
4109 fprintf(file, " %s ", newdelete);
4110 write_name_quoted(fs->path, file, '\n');
4111 }
4112
4113
4114 static void show_mode_change(FILE *file, struct diff_filepair *p, int show_name,
4115 const char *line_prefix)
4116 {
4117 if (p->one->mode && p->two->mode && p->one->mode != p->two->mode) {
4118 fprintf(file, "%s mode change %06o => %06o%c", line_prefix, p->one->mode,
4119 p->two->mode, show_name ? ' ' : '\n');
4120 if (show_name) {
4121 write_name_quoted(p->two->path, file, '\n');
4122 }
4123 }
4124 }
4125
4126 static void show_rename_copy(FILE *file, const char *renamecopy, struct diff_filepair *p,
4127 const char *line_prefix)
4128 {
4129 char *names = pprint_rename(p->one->path, p->two->path);
4130
4131 fprintf(file, " %s %s (%d%%)\n", renamecopy, names, similarity_index(p));
4132 free(names);
4133 show_mode_change(file, p, 0, line_prefix);
4134 }
4135
4136 static void diff_summary(struct diff_options *opt, struct diff_filepair *p)
4137 {
4138 FILE *file = opt->file;
4139 char *line_prefix = "";
4140
4141 if (opt->output_prefix) {
4142 struct strbuf *buf = opt->output_prefix(opt, opt->output_prefix_data);
4143 line_prefix = buf->buf;
4144 }
4145
4146 switch(p->status) {
4147 case DIFF_STATUS_DELETED:
4148 fputs(line_prefix, file);
4149 show_file_mode_name(file, "delete", p->one);
4150 break;
4151 case DIFF_STATUS_ADDED:
4152 fputs(line_prefix, file);
4153 show_file_mode_name(file, "create", p->two);
4154 break;
4155 case DIFF_STATUS_COPIED:
4156 fputs(line_prefix, file);
4157 show_rename_copy(file, "copy", p, line_prefix);
4158 break;
4159 case DIFF_STATUS_RENAMED:
4160 fputs(line_prefix, file);
4161 show_rename_copy(file, "rename", p, line_prefix);
4162 break;
4163 default:
4164 if (p->score) {
4165 fprintf(file, "%s rewrite ", line_prefix);
4166 write_name_quoted(p->two->path, file, ' ');
4167 fprintf(file, "(%d%%)\n", similarity_index(p));
4168 }
4169 show_mode_change(file, p, !p->score, line_prefix);
4170 break;
4171 }
4172 }
4173
4174 struct patch_id_t {
4175 git_SHA_CTX *ctx;
4176 int patchlen;
4177 };
4178
4179 static int remove_space(char *line, int len)
4180 {
4181 int i;
4182 char *dst = line;
4183 unsigned char c;
4184
4185 for (i = 0; i < len; i++)
4186 if (!isspace((c = line[i])))
4187 *dst++ = c;
4188
4189 return dst - line;
4190 }
4191
4192 static void patch_id_consume(void *priv, char *line, unsigned long len)
4193 {
4194 struct patch_id_t *data = priv;
4195 int new_len;
4196
4197 /* Ignore line numbers when computing the SHA1 of the patch */
4198 if (!prefixcmp(line, "@@ -"))
4199 return;
4200
4201 new_len = remove_space(line, len);
4202
4203 git_SHA1_Update(data->ctx, line, new_len);
4204 data->patchlen += new_len;
4205 }
4206
4207 /* returns 0 upon success, and writes result into sha1 */
4208 static int diff_get_patch_id(struct diff_options *options, unsigned char *sha1)
4209 {
4210 struct diff_queue_struct *q = &diff_queued_diff;
4211 int i;
4212 git_SHA_CTX ctx;
4213 struct patch_id_t data;
4214 char buffer[PATH_MAX * 4 + 20];
4215
4216 git_SHA1_Init(&ctx);
4217 memset(&data, 0, sizeof(struct patch_id_t));
4218 data.ctx = &ctx;
4219
4220 for (i = 0; i < q->nr; i++) {
4221 xpparam_t xpp;
4222 xdemitconf_t xecfg;
4223 mmfile_t mf1, mf2;
4224 struct diff_filepair *p = q->queue[i];
4225 int len1, len2;
4226
4227 memset(&xpp, 0, sizeof(xpp));
4228 memset(&xecfg, 0, sizeof(xecfg));
4229 if (p->status == 0)
4230 return error("internal diff status error");
4231 if (p->status == DIFF_STATUS_UNKNOWN)
4232 continue;
4233 if (diff_unmodified_pair(p))
4234 continue;
4235 if ((DIFF_FILE_VALID(p->one) && S_ISDIR(p->one->mode)) ||
4236 (DIFF_FILE_VALID(p->two) && S_ISDIR(p->two->mode)))
4237 continue;
4238 if (DIFF_PAIR_UNMERGED(p))
4239 continue;
4240
4241 diff_fill_sha1_info(p->one);
4242 diff_fill_sha1_info(p->two);
4243 if (fill_mmfile(&mf1, p->one) < 0 ||
4244 fill_mmfile(&mf2, p->two) < 0)
4245 return error("unable to read files to diff");
4246
4247 len1 = remove_space(p->one->path, strlen(p->one->path));
4248 len2 = remove_space(p->two->path, strlen(p->two->path));
4249 if (p->one->mode == 0)
4250 len1 = snprintf(buffer, sizeof(buffer),
4251 "diff--gita/%.*sb/%.*s"
4252 "newfilemode%06o"
4253 "---/dev/null"
4254 "+++b/%.*s",
4255 len1, p->one->path,
4256 len2, p->two->path,
4257 p->two->mode,
4258 len2, p->two->path);
4259 else if (p->two->mode == 0)
4260 len1 = snprintf(buffer, sizeof(buffer),
4261 "diff--gita/%.*sb/%.*s"
4262 "deletedfilemode%06o"
4263 "---a/%.*s"
4264 "+++/dev/null",
4265 len1, p->one->path,
4266 len2, p->two->path,
4267 p->one->mode,
4268 len1, p->one->path);
4269 else
4270 len1 = snprintf(buffer, sizeof(buffer),
4271 "diff--gita/%.*sb/%.*s"
4272 "---a/%.*s"
4273 "+++b/%.*s",
4274 len1, p->one->path,
4275 len2, p->two->path,
4276 len1, p->one->path,
4277 len2, p->two->path);
4278 git_SHA1_Update(&ctx, buffer, len1);
4279
4280 if (diff_filespec_is_binary(p->one) ||
4281 diff_filespec_is_binary(p->two)) {
4282 git_SHA1_Update(&ctx, sha1_to_hex(p->one->sha1), 40);
4283 git_SHA1_Update(&ctx, sha1_to_hex(p->two->sha1), 40);
4284 continue;
4285 }
4286
4287 xpp.flags = 0;
4288 xecfg.ctxlen = 3;
4289 xecfg.flags = 0;
4290 xdi_diff_outf(&mf1, &mf2, patch_id_consume, &data,
4291 &xpp, &xecfg);
4292 }
4293
4294 git_SHA1_Final(sha1, &ctx);
4295 return 0;
4296 }
4297
4298 int diff_flush_patch_id(struct diff_options *options, unsigned char *sha1)
4299 {
4300 struct diff_queue_struct *q = &diff_queued_diff;
4301 int i;
4302 int result = diff_get_patch_id(options, sha1);
4303
4304 for (i = 0; i < q->nr; i++)
4305 diff_free_filepair(q->queue[i]);
4306
4307 free(q->queue);
4308 DIFF_QUEUE_CLEAR(q);
4309
4310 return result;
4311 }
4312
4313 static int is_summary_empty(const struct diff_queue_struct *q)
4314 {
4315 int i;
4316
4317 for (i = 0; i < q->nr; i++) {
4318 const struct diff_filepair *p = q->queue[i];
4319
4320 switch (p->status) {
4321 case DIFF_STATUS_DELETED:
4322 case DIFF_STATUS_ADDED:
4323 case DIFF_STATUS_COPIED:
4324 case DIFF_STATUS_RENAMED:
4325 return 0;
4326 default:
4327 if (p->score)
4328 return 0;
4329 if (p->one->mode && p->two->mode &&
4330 p->one->mode != p->two->mode)
4331 return 0;
4332 break;
4333 }
4334 }
4335 return 1;
4336 }
4337
4338 static const char rename_limit_warning[] =
4339 "inexact rename detection was skipped due to too many files.";
4340
4341 static const char degrade_cc_to_c_warning[] =
4342 "only found copies from modified paths due to too many files.";
4343
4344 static const char rename_limit_advice[] =
4345 "you may want to set your %s variable to at least "
4346 "%d and retry the command.";
4347
4348 void diff_warn_rename_limit(const char *varname, int needed, int degraded_cc)
4349 {
4350 if (degraded_cc)
4351 warning(degrade_cc_to_c_warning);
4352 else if (needed)
4353 warning(rename_limit_warning);
4354 else
4355 return;
4356 if (0 < needed && needed < 32767)
4357 warning(rename_limit_advice, varname, needed);
4358 }
4359
4360 void diff_flush(struct diff_options *options)
4361 {
4362 struct diff_queue_struct *q = &diff_queued_diff;
4363 int i, output_format = options->output_format;
4364 int separator = 0;
4365 int dirstat_by_line = 0;
4366
4367 /*
4368 * Order: raw, stat, summary, patch
4369 * or: name/name-status/checkdiff (other bits clear)
4370 */
4371 if (!q->nr)
4372 goto free_queue;
4373
4374 if (output_format & (DIFF_FORMAT_RAW |
4375 DIFF_FORMAT_NAME |
4376 DIFF_FORMAT_NAME_STATUS |
4377 DIFF_FORMAT_CHECKDIFF)) {
4378 for (i = 0; i < q->nr; i++) {
4379 struct diff_filepair *p = q->queue[i];
4380 if (check_pair_status(p))
4381 flush_one_pair(p, options);
4382 }
4383 separator++;
4384 }
4385
4386 if (output_format & DIFF_FORMAT_DIRSTAT && DIFF_OPT_TST(options, DIRSTAT_BY_LINE))
4387 dirstat_by_line = 1;
4388
4389 if (output_format & (DIFF_FORMAT_DIFFSTAT|DIFF_FORMAT_SHORTSTAT|DIFF_FORMAT_NUMSTAT) ||
4390 dirstat_by_line) {
4391 struct diffstat_t diffstat;
4392
4393 memset(&diffstat, 0, sizeof(struct diffstat_t));
4394 for (i = 0; i < q->nr; i++) {
4395 struct diff_filepair *p = q->queue[i];
4396 if (check_pair_status(p))
4397 diff_flush_stat(p, options, &diffstat);
4398 }
4399 if (output_format & DIFF_FORMAT_NUMSTAT)
4400 show_numstat(&diffstat, options);
4401 if (output_format & DIFF_FORMAT_DIFFSTAT)
4402 show_stats(&diffstat, options);
4403 if (output_format & DIFF_FORMAT_SHORTSTAT)
4404 show_shortstats(&diffstat, options);
4405 if (output_format & DIFF_FORMAT_DIRSTAT)
4406 show_dirstat_by_line(&diffstat, options);
4407 free_diffstat_info(&diffstat);
4408 separator++;
4409 }
4410 if ((output_format & DIFF_FORMAT_DIRSTAT) && !dirstat_by_line)
4411 show_dirstat(options);
4412
4413 if (output_format & DIFF_FORMAT_SUMMARY && !is_summary_empty(q)) {
4414 for (i = 0; i < q->nr; i++) {
4415 diff_summary(options, q->queue[i]);
4416 }
4417 separator++;
4418 }
4419
4420 if (output_format & DIFF_FORMAT_NO_OUTPUT &&
4421 DIFF_OPT_TST(options, EXIT_WITH_STATUS) &&
4422 DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4423 /*
4424 * run diff_flush_patch for the exit status. setting
4425 * options->file to /dev/null should be safe, becaue we
4426 * aren't supposed to produce any output anyway.
4427 */
4428 if (options->close_file)
4429 fclose(options->file);
4430 options->file = fopen("/dev/null", "w");
4431 if (!options->file)
4432 die_errno("Could not open /dev/null");
4433 options->close_file = 1;
4434 for (i = 0; i < q->nr; i++) {
4435 struct diff_filepair *p = q->queue[i];
4436 if (check_pair_status(p))
4437 diff_flush_patch(p, options);
4438 if (options->found_changes)
4439 break;
4440 }
4441 }
4442
4443 if (output_format & DIFF_FORMAT_PATCH) {
4444 if (separator) {
4445 if (options->output_prefix) {
4446 struct strbuf *msg = NULL;
4447 msg = options->output_prefix(options,
4448 options->output_prefix_data);
4449 fwrite(msg->buf, msg->len, 1, stdout);
4450 }
4451 putc(options->line_termination, options->file);
4452 if (options->stat_sep) {
4453 /* attach patch instead of inline */
4454 fputs(options->stat_sep, options->file);
4455 }
4456 }
4457
4458 for (i = 0; i < q->nr; i++) {
4459 struct diff_filepair *p = q->queue[i];
4460 if (check_pair_status(p))
4461 diff_flush_patch(p, options);
4462 }
4463 }
4464
4465 if (output_format & DIFF_FORMAT_CALLBACK)
4466 options->format_callback(q, options, options->format_callback_data);
4467
4468 for (i = 0; i < q->nr; i++)
4469 diff_free_filepair(q->queue[i]);
4470 free_queue:
4471 free(q->queue);
4472 DIFF_QUEUE_CLEAR(q);
4473 if (options->close_file)
4474 fclose(options->file);
4475
4476 /*
4477 * Report the content-level differences with HAS_CHANGES;
4478 * diff_addremove/diff_change does not set the bit when
4479 * DIFF_FROM_CONTENTS is in effect (e.g. with -w).
4480 */
4481 if (DIFF_OPT_TST(options, DIFF_FROM_CONTENTS)) {
4482 if (options->found_changes)
4483 DIFF_OPT_SET(options, HAS_CHANGES);
4484 else
4485 DIFF_OPT_CLR(options, HAS_CHANGES);
4486 }
4487 }
4488
4489 static void diffcore_apply_filter(const char *filter)
4490 {
4491 int i;
4492 struct diff_queue_struct *q = &diff_queued_diff;
4493 struct diff_queue_struct outq;
4494 DIFF_QUEUE_CLEAR(&outq);
4495
4496 if (!filter)
4497 return;
4498
4499 if (strchr(filter, DIFF_STATUS_FILTER_AON)) {
4500 int found;
4501 for (i = found = 0; !found && i < q->nr; i++) {
4502 struct diff_filepair *p = q->queue[i];
4503 if (((p->status == DIFF_STATUS_MODIFIED) &&
4504 ((p->score &&
4505 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4506 (!p->score &&
4507 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4508 ((p->status != DIFF_STATUS_MODIFIED) &&
4509 strchr(filter, p->status)))
4510 found++;
4511 }
4512 if (found)
4513 return;
4514
4515 /* otherwise we will clear the whole queue
4516 * by copying the empty outq at the end of this
4517 * function, but first clear the current entries
4518 * in the queue.
4519 */
4520 for (i = 0; i < q->nr; i++)
4521 diff_free_filepair(q->queue[i]);
4522 }
4523 else {
4524 /* Only the matching ones */
4525 for (i = 0; i < q->nr; i++) {
4526 struct diff_filepair *p = q->queue[i];
4527
4528 if (((p->status == DIFF_STATUS_MODIFIED) &&
4529 ((p->score &&
4530 strchr(filter, DIFF_STATUS_FILTER_BROKEN)) ||
4531 (!p->score &&
4532 strchr(filter, DIFF_STATUS_MODIFIED)))) ||
4533 ((p->status != DIFF_STATUS_MODIFIED) &&
4534 strchr(filter, p->status)))
4535 diff_q(&outq, p);
4536 else
4537 diff_free_filepair(p);
4538 }
4539 }
4540 free(q->queue);
4541 *q = outq;
4542 }
4543
4544 /* Check whether two filespecs with the same mode and size are identical */
4545 static int diff_filespec_is_identical(struct diff_filespec *one,
4546 struct diff_filespec *two)
4547 {
4548 if (S_ISGITLINK(one->mode))
4549 return 0;
4550 if (diff_populate_filespec(one, 0))
4551 return 0;
4552 if (diff_populate_filespec(two, 0))
4553 return 0;
4554 return !memcmp(one->data, two->data, one->size);
4555 }
4556
4557 static void diffcore_skip_stat_unmatch(struct diff_options *diffopt)
4558 {
4559 int i;
4560 struct diff_queue_struct *q = &diff_queued_diff;
4561 struct diff_queue_struct outq;
4562 DIFF_QUEUE_CLEAR(&outq);
4563
4564 for (i = 0; i < q->nr; i++) {
4565 struct diff_filepair *p = q->queue[i];
4566
4567 /*
4568 * 1. Entries that come from stat info dirtiness
4569 * always have both sides (iow, not create/delete),
4570 * one side of the object name is unknown, with
4571 * the same mode and size. Keep the ones that
4572 * do not match these criteria. They have real
4573 * differences.
4574 *
4575 * 2. At this point, the file is known to be modified,
4576 * with the same mode and size, and the object
4577 * name of one side is unknown. Need to inspect
4578 * the identical contents.
4579 */
4580 if (!DIFF_FILE_VALID(p->one) || /* (1) */
4581 !DIFF_FILE_VALID(p->two) ||
4582 (p->one->sha1_valid && p->two->sha1_valid) ||
4583 (p->one->mode != p->two->mode) ||
4584 diff_populate_filespec(p->one, 1) ||
4585 diff_populate_filespec(p->two, 1) ||
4586 (p->one->size != p->two->size) ||
4587 !diff_filespec_is_identical(p->one, p->two)) /* (2) */
4588 diff_q(&outq, p);
4589 else {
4590 /*
4591 * The caller can subtract 1 from skip_stat_unmatch
4592 * to determine how many paths were dirty only
4593 * due to stat info mismatch.
4594 */
4595 if (!DIFF_OPT_TST(diffopt, NO_INDEX))
4596 diffopt->skip_stat_unmatch++;
4597 diff_free_filepair(p);
4598 }
4599 }
4600 free(q->queue);
4601 *q = outq;
4602 }
4603
4604 static int diffnamecmp(const void *a_, const void *b_)
4605 {
4606 const struct diff_filepair *a = *((const struct diff_filepair **)a_);
4607 const struct diff_filepair *b = *((const struct diff_filepair **)b_);
4608 const char *name_a, *name_b;
4609
4610 name_a = a->one ? a->one->path : a->two->path;
4611 name_b = b->one ? b->one->path : b->two->path;
4612 return strcmp(name_a, name_b);
4613 }
4614
4615 void diffcore_fix_diff_index(struct diff_options *options)
4616 {
4617 struct diff_queue_struct *q = &diff_queued_diff;
4618 qsort(q->queue, q->nr, sizeof(q->queue[0]), diffnamecmp);
4619 }
4620
4621 void diffcore_std(struct diff_options *options)
4622 {
4623 if (options->skip_stat_unmatch)
4624 diffcore_skip_stat_unmatch(options);
4625 if (!options->found_follow) {
4626 /* See try_to_follow_renames() in tree-diff.c */
4627 if (options->break_opt != -1)
4628 diffcore_break(options->break_opt);
4629 if (options->detect_rename)
4630 diffcore_rename(options);
4631 if (options->break_opt != -1)
4632 diffcore_merge_broken();
4633 }
4634 if (options->pickaxe)
4635 diffcore_pickaxe(options);
4636 if (options->orderfile)
4637 diffcore_order(options->orderfile);
4638 if (!options->found_follow)
4639 /* See try_to_follow_renames() in tree-diff.c */
4640 diff_resolve_rename_copy();
4641 diffcore_apply_filter(options->filter);
4642
4643 if (diff_queued_diff.nr && !DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4644 DIFF_OPT_SET(options, HAS_CHANGES);
4645 else
4646 DIFF_OPT_CLR(options, HAS_CHANGES);
4647
4648 options->found_follow = 0;
4649 }
4650
4651 int diff_result_code(struct diff_options *opt, int status)
4652 {
4653 int result = 0;
4654
4655 diff_warn_rename_limit("diff.renamelimit",
4656 opt->needed_rename_limit,
4657 opt->degraded_cc_to_c);
4658 if (!DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4659 !(opt->output_format & DIFF_FORMAT_CHECKDIFF))
4660 return status;
4661 if (DIFF_OPT_TST(opt, EXIT_WITH_STATUS) &&
4662 DIFF_OPT_TST(opt, HAS_CHANGES))
4663 result |= 01;
4664 if ((opt->output_format & DIFF_FORMAT_CHECKDIFF) &&
4665 DIFF_OPT_TST(opt, CHECK_FAILED))
4666 result |= 02;
4667 return result;
4668 }
4669
4670 int diff_can_quit_early(struct diff_options *opt)
4671 {
4672 return (DIFF_OPT_TST(opt, QUICK) &&
4673 !opt->filter &&
4674 DIFF_OPT_TST(opt, HAS_CHANGES));
4675 }
4676
4677 /*
4678 * Shall changes to this submodule be ignored?
4679 *
4680 * Submodule changes can be configured to be ignored separately for each path,
4681 * but that configuration can be overridden from the command line.
4682 */
4683 static int is_submodule_ignored(const char *path, struct diff_options *options)
4684 {
4685 int ignored = 0;
4686 unsigned orig_flags = options->flags;
4687 if (!DIFF_OPT_TST(options, OVERRIDE_SUBMODULE_CONFIG))
4688 set_diffopt_flags_from_submodule_config(options, path);
4689 if (DIFF_OPT_TST(options, IGNORE_SUBMODULES))
4690 ignored = 1;
4691 options->flags = orig_flags;
4692 return ignored;
4693 }
4694
4695 void diff_addremove(struct diff_options *options,
4696 int addremove, unsigned mode,
4697 const unsigned char *sha1,
4698 const char *concatpath, unsigned dirty_submodule)
4699 {
4700 struct diff_filespec *one, *two;
4701
4702 if (S_ISGITLINK(mode) && is_submodule_ignored(concatpath, options))
4703 return;
4704
4705 /* This may look odd, but it is a preparation for
4706 * feeding "there are unchanged files which should
4707 * not produce diffs, but when you are doing copy
4708 * detection you would need them, so here they are"
4709 * entries to the diff-core. They will be prefixed
4710 * with something like '=' or '*' (I haven't decided
4711 * which but should not make any difference).
4712 * Feeding the same new and old to diff_change()
4713 * also has the same effect.
4714 * Before the final output happens, they are pruned after
4715 * merged into rename/copy pairs as appropriate.
4716 */
4717 if (DIFF_OPT_TST(options, REVERSE_DIFF))
4718 addremove = (addremove == '+' ? '-' :
4719 addremove == '-' ? '+' : addremove);
4720
4721 if (options->prefix &&
4722 strncmp(concatpath, options->prefix, options->prefix_length))
4723 return;
4724
4725 one = alloc_filespec(concatpath);
4726 two = alloc_filespec(concatpath);
4727
4728 if (addremove != '+')
4729 fill_filespec(one, sha1, mode);
4730 if (addremove != '-') {
4731 fill_filespec(two, sha1, mode);
4732 two->dirty_submodule = dirty_submodule;
4733 }
4734
4735 diff_queue(&diff_queued_diff, one, two);
4736 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4737 DIFF_OPT_SET(options, HAS_CHANGES);
4738 }
4739
4740 void diff_change(struct diff_options *options,
4741 unsigned old_mode, unsigned new_mode,
4742 const unsigned char *old_sha1,
4743 const unsigned char *new_sha1,
4744 const char *concatpath,
4745 unsigned old_dirty_submodule, unsigned new_dirty_submodule)
4746 {
4747 struct diff_filespec *one, *two;
4748
4749 if (S_ISGITLINK(old_mode) && S_ISGITLINK(new_mode) &&
4750 is_submodule_ignored(concatpath, options))
4751 return;
4752
4753 if (DIFF_OPT_TST(options, REVERSE_DIFF)) {
4754 unsigned tmp;
4755 const unsigned char *tmp_c;
4756 tmp = old_mode; old_mode = new_mode; new_mode = tmp;
4757 tmp_c = old_sha1; old_sha1 = new_sha1; new_sha1 = tmp_c;
4758 tmp = old_dirty_submodule; old_dirty_submodule = new_dirty_submodule;
4759 new_dirty_submodule = tmp;
4760 }
4761
4762 if (options->prefix &&
4763 strncmp(concatpath, options->prefix, options->prefix_length))
4764 return;
4765
4766 one = alloc_filespec(concatpath);
4767 two = alloc_filespec(concatpath);
4768 fill_filespec(one, old_sha1, old_mode);
4769 fill_filespec(two, new_sha1, new_mode);
4770 one->dirty_submodule = old_dirty_submodule;
4771 two->dirty_submodule = new_dirty_submodule;
4772
4773 diff_queue(&diff_queued_diff, one, two);
4774 if (!DIFF_OPT_TST(options, DIFF_FROM_CONTENTS))
4775 DIFF_OPT_SET(options, HAS_CHANGES);
4776 }
4777
4778 struct diff_filepair *diff_unmerge(struct diff_options *options, const char *path)
4779 {
4780 struct diff_filepair *pair;
4781 struct diff_filespec *one, *two;
4782
4783 if (options->prefix &&
4784 strncmp(path, options->prefix, options->prefix_length))
4785 return NULL;
4786
4787 one = alloc_filespec(path);
4788 two = alloc_filespec(path);
4789 pair = diff_queue(&diff_queued_diff, one, two);
4790 pair->is_unmerged = 1;
4791 return pair;
4792 }
4793
4794 static char *run_textconv(const char *pgm, struct diff_filespec *spec,
4795 size_t *outsize)
4796 {
4797 struct diff_tempfile *temp;
4798 const char *argv[3];
4799 const char **arg = argv;
4800 struct child_process child;
4801 struct strbuf buf = STRBUF_INIT;
4802 int err = 0;
4803
4804 temp = prepare_temp_file(spec->path, spec);
4805 *arg++ = pgm;
4806 *arg++ = temp->name;
4807 *arg = NULL;
4808
4809 memset(&child, 0, sizeof(child));
4810 child.use_shell = 1;
4811 child.argv = argv;
4812 child.out = -1;
4813 if (start_command(&child)) {
4814 remove_tempfile();
4815 return NULL;
4816 }
4817
4818 if (strbuf_read(&buf, child.out, 0) < 0)
4819 err = error("error reading from textconv command '%s'", pgm);
4820 close(child.out);
4821
4822 if (finish_command(&child) || err) {
4823 strbuf_release(&buf);
4824 remove_tempfile();
4825 return NULL;
4826 }
4827 remove_tempfile();
4828
4829 return strbuf_detach(&buf, outsize);
4830 }
4831
4832 size_t fill_textconv(struct userdiff_driver *driver,
4833 struct diff_filespec *df,
4834 char **outbuf)
4835 {
4836 size_t size;
4837
4838 if (!driver || !driver->textconv) {
4839 if (!DIFF_FILE_VALID(df)) {
4840 *outbuf = "";
4841 return 0;
4842 }
4843 if (diff_populate_filespec(df, 0))
4844 die("unable to read files to diff");
4845 *outbuf = df->data;
4846 return df->size;
4847 }
4848
4849 if (driver->textconv_cache && df->sha1_valid) {
4850 *outbuf = notes_cache_get(driver->textconv_cache, df->sha1,
4851 &size);
4852 if (*outbuf)
4853 return size;
4854 }
4855
4856 *outbuf = run_textconv(driver->textconv, df, &size);
4857 if (!*outbuf)
4858 die("unable to read files to diff");
4859
4860 if (driver->textconv_cache && df->sha1_valid) {
4861 /* ignore errors, as we might be in a readonly repository */
4862 notes_cache_put(driver->textconv_cache, df->sha1, *outbuf,
4863 size);
4864 /*
4865 * we could save up changes and flush them all at the end,
4866 * but we would need an extra call after all diffing is done.
4867 * Since generating a cache entry is the slow path anyway,
4868 * this extra overhead probably isn't a big deal.
4869 */
4870 notes_cache_write(driver->textconv_cache);
4871 }
4872
4873 return size;
4874 }
git with Jakub Narębski modifications