Documentation for git-fmt-merge-msg
[git/dscho.git] / apply.c
blob3e53b3438169bcaaf4db97669b062ebe14335fe6
1 /*
2 * apply.c
4 * Copyright (C) Linus Torvalds, 2005
6 * This applies patches on top of some (arbitrary) version of the SCM.
8 */
9 #include <fnmatch.h>
10 #include "cache.h"
11 #include "quote.h"
13 // --check turns on checking that the working tree matches the
14 // files that are being modified, but doesn't apply the patch
15 // --stat does just a diffstat, and doesn't actually apply
16 // --numstat does numeric diffstat, and doesn't actually apply
17 // --index-info shows the old and new index info for paths if available.
19 static int check_index = 0;
20 static int write_index = 0;
21 static int diffstat = 0;
22 static int numstat = 0;
23 static int summary = 0;
24 static int check = 0;
25 static int apply = 1;
26 static int show_index_info = 0;
27 static int line_termination = '\n';
28 static const char apply_usage[] =
29 "git-apply [--stat] [--numstat] [--summary] [--check] [--index] [--apply] [--index-info] [-z] <patch>...";
32 * For "diff-stat" like behaviour, we keep track of the biggest change
33 * we've seen, and the longest filename. That allows us to do simple
34 * scaling.
36 static int max_change, max_len;
39 * Various "current state", notably line numbers and what
40 * file (and how) we're patching right now.. The "is_xxxx"
41 * things are flags, where -1 means "don't know yet".
43 static int linenr = 1;
45 struct fragment {
46 unsigned long oldpos, oldlines;
47 unsigned long newpos, newlines;
48 const char *patch;
49 int size;
50 struct fragment *next;
53 struct patch {
54 char *new_name, *old_name, *def_name;
55 unsigned int old_mode, new_mode;
56 int is_rename, is_copy, is_new, is_delete;
57 int lines_added, lines_deleted;
58 int score;
59 struct fragment *fragments;
60 char *result;
61 unsigned long resultsize;
62 char old_sha1_prefix[41];
63 char new_sha1_prefix[41];
64 struct patch *next;
67 #define CHUNKSIZE (8192)
68 #define SLOP (16)
70 static void *read_patch_file(int fd, unsigned long *sizep)
72 unsigned long size = 0, alloc = CHUNKSIZE;
73 void *buffer = xmalloc(alloc);
75 for (;;) {
76 int nr = alloc - size;
77 if (nr < 1024) {
78 alloc += CHUNKSIZE;
79 buffer = xrealloc(buffer, alloc);
80 nr = alloc - size;
82 nr = read(fd, buffer + size, nr);
83 if (!nr)
84 break;
85 if (nr < 0) {
86 if (errno == EAGAIN)
87 continue;
88 die("git-apply: read returned %s", strerror(errno));
90 size += nr;
92 *sizep = size;
95 * Make sure that we have some slop in the buffer
96 * so that we can do speculative "memcmp" etc, and
97 * see to it that it is NUL-filled.
99 if (alloc < size + SLOP)
100 buffer = xrealloc(buffer, size + SLOP);
101 memset(buffer + size, 0, SLOP);
102 return buffer;
105 static unsigned long linelen(const char *buffer, unsigned long size)
107 unsigned long len = 0;
108 while (size--) {
109 len++;
110 if (*buffer++ == '\n')
111 break;
113 return len;
116 static int is_dev_null(const char *str)
118 return !memcmp("/dev/null", str, 9) && isspace(str[9]);
121 #define TERM_SPACE 1
122 #define TERM_TAB 2
124 static int name_terminate(const char *name, int namelen, int c, int terminate)
126 if (c == ' ' && !(terminate & TERM_SPACE))
127 return 0;
128 if (c == '\t' && !(terminate & TERM_TAB))
129 return 0;
131 return 1;
134 static char * find_name(const char *line, char *def, int p_value, int terminate)
136 int len;
137 const char *start = line;
138 char *name;
140 if (*line == '"') {
141 /* Proposed "new-style" GNU patch/diff format; see
142 * http://marc.theaimsgroup.com/?l=git&m=112927316408690&w=2
144 name = unquote_c_style(line, NULL);
145 if (name) {
146 char *cp = name;
147 while (p_value) {
148 cp = strchr(name, '/');
149 if (!cp)
150 break;
151 cp++;
152 p_value--;
154 if (cp) {
155 /* name can later be freed, so we need
156 * to memmove, not just return cp
158 memmove(name, cp, strlen(cp) + 1);
159 free(def);
160 return name;
162 else {
163 free(name);
164 name = NULL;
169 for (;;) {
170 char c = *line;
172 if (isspace(c)) {
173 if (c == '\n')
174 break;
175 if (name_terminate(start, line-start, c, terminate))
176 break;
178 line++;
179 if (c == '/' && !--p_value)
180 start = line;
182 if (!start)
183 return def;
184 len = line - start;
185 if (!len)
186 return def;
189 * Generally we prefer the shorter name, especially
190 * if the other one is just a variation of that with
191 * something else tacked on to the end (ie "file.orig"
192 * or "file~").
194 if (def) {
195 int deflen = strlen(def);
196 if (deflen < len && !strncmp(start, def, deflen))
197 return def;
200 name = xmalloc(len + 1);
201 memcpy(name, start, len);
202 name[len] = 0;
203 free(def);
204 return name;
208 * Get the name etc info from the --/+++ lines of a traditional patch header
210 * NOTE! This hardcodes "-p1" behaviour in filename detection.
212 * FIXME! The end-of-filename heuristics are kind of screwy. For existing
213 * files, we can happily check the index for a match, but for creating a
214 * new file we should try to match whatever "patch" does. I have no idea.
216 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
218 int p_value = 1;
219 char *name;
221 first += 4; // skip "--- "
222 second += 4; // skip "+++ "
223 if (is_dev_null(first)) {
224 patch->is_new = 1;
225 patch->is_delete = 0;
226 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
227 patch->new_name = name;
228 } else if (is_dev_null(second)) {
229 patch->is_new = 0;
230 patch->is_delete = 1;
231 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
232 patch->old_name = name;
233 } else {
234 name = find_name(first, NULL, p_value, TERM_SPACE | TERM_TAB);
235 name = find_name(second, name, p_value, TERM_SPACE | TERM_TAB);
236 patch->old_name = patch->new_name = name;
238 if (!name)
239 die("unable to find filename in patch at line %d", linenr);
242 static int gitdiff_hdrend(const char *line, struct patch *patch)
244 return -1;
248 * We're anal about diff header consistency, to make
249 * sure that we don't end up having strange ambiguous
250 * patches floating around.
252 * As a result, gitdiff_{old|new}name() will check
253 * their names against any previous information, just
254 * to make sure..
256 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
258 if (!orig_name && !isnull)
259 return find_name(line, NULL, 1, 0);
261 if (orig_name) {
262 int len;
263 const char *name;
264 char *another;
265 name = orig_name;
266 len = strlen(name);
267 if (isnull)
268 die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
269 another = find_name(line, NULL, 1, 0);
270 if (!another || memcmp(another, name, len))
271 die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
272 free(another);
273 return orig_name;
275 else {
276 /* expect "/dev/null" */
277 if (memcmp("/dev/null", line, 9) || line[9] != '\n')
278 die("git-apply: bad git-diff - expected /dev/null on line %d", linenr);
279 return NULL;
283 static int gitdiff_oldname(const char *line, struct patch *patch)
285 patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
286 return 0;
289 static int gitdiff_newname(const char *line, struct patch *patch)
291 patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
292 return 0;
295 static int gitdiff_oldmode(const char *line, struct patch *patch)
297 patch->old_mode = strtoul(line, NULL, 8);
298 return 0;
301 static int gitdiff_newmode(const char *line, struct patch *patch)
303 patch->new_mode = strtoul(line, NULL, 8);
304 return 0;
307 static int gitdiff_delete(const char *line, struct patch *patch)
309 patch->is_delete = 1;
310 patch->old_name = patch->def_name;
311 return gitdiff_oldmode(line, patch);
314 static int gitdiff_newfile(const char *line, struct patch *patch)
316 patch->is_new = 1;
317 patch->new_name = patch->def_name;
318 return gitdiff_newmode(line, patch);
321 static int gitdiff_copysrc(const char *line, struct patch *patch)
323 patch->is_copy = 1;
324 patch->old_name = find_name(line, NULL, 0, 0);
325 return 0;
328 static int gitdiff_copydst(const char *line, struct patch *patch)
330 patch->is_copy = 1;
331 patch->new_name = find_name(line, NULL, 0, 0);
332 return 0;
335 static int gitdiff_renamesrc(const char *line, struct patch *patch)
337 patch->is_rename = 1;
338 patch->old_name = find_name(line, NULL, 0, 0);
339 return 0;
342 static int gitdiff_renamedst(const char *line, struct patch *patch)
344 patch->is_rename = 1;
345 patch->new_name = find_name(line, NULL, 0, 0);
346 return 0;
349 static int gitdiff_similarity(const char *line, struct patch *patch)
351 if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
352 patch->score = 0;
353 return 0;
356 static int gitdiff_dissimilarity(const char *line, struct patch *patch)
358 if ((patch->score = strtoul(line, NULL, 10)) == ULONG_MAX)
359 patch->score = 0;
360 return 0;
363 static int gitdiff_index(const char *line, struct patch *patch)
365 /* index line is N hexadecimal, "..", N hexadecimal,
366 * and optional space with octal mode.
368 const char *ptr, *eol;
369 int len;
371 ptr = strchr(line, '.');
372 if (!ptr || ptr[1] != '.' || 40 <= ptr - line)
373 return 0;
374 len = ptr - line;
375 memcpy(patch->old_sha1_prefix, line, len);
376 patch->old_sha1_prefix[len] = 0;
378 line = ptr + 2;
379 ptr = strchr(line, ' ');
380 eol = strchr(line, '\n');
382 if (!ptr || eol < ptr)
383 ptr = eol;
384 len = ptr - line;
386 if (40 <= len)
387 return 0;
388 memcpy(patch->new_sha1_prefix, line, len);
389 patch->new_sha1_prefix[len] = 0;
390 if (*ptr == ' ')
391 patch->new_mode = patch->old_mode = strtoul(ptr+1, NULL, 8);
392 return 0;
396 * This is normal for a diff that doesn't change anything: we'll fall through
397 * into the next diff. Tell the parser to break out.
399 static int gitdiff_unrecognized(const char *line, struct patch *patch)
401 return -1;
404 static const char *stop_at_slash(const char *line, int llen)
406 int i;
408 for (i = 0; i < llen; i++) {
409 int ch = line[i];
410 if (ch == '/')
411 return line + i;
413 return NULL;
416 /* This is to extract the same name that appears on "diff --git"
417 * line. We do not find and return anything if it is a rename
418 * patch, and it is OK because we will find the name elsewhere.
419 * We need to reliably find name only when it is mode-change only,
420 * creation or deletion of an empty file. In any of these cases,
421 * both sides are the same name under a/ and b/ respectively.
423 static char *git_header_name(char *line, int llen)
425 int len;
426 const char *name;
427 const char *second = NULL;
429 line += strlen("diff --git ");
430 llen -= strlen("diff --git ");
432 if (*line == '"') {
433 const char *cp;
434 char *first = unquote_c_style(line, &second);
435 if (!first)
436 return NULL;
438 /* advance to the first slash */
439 cp = stop_at_slash(first, strlen(first));
440 if (!cp || cp == first) {
441 /* we do not accept absolute paths */
442 free_first_and_fail:
443 free(first);
444 return NULL;
446 len = strlen(cp+1);
447 memmove(first, cp+1, len+1); /* including NUL */
449 /* second points at one past closing dq of name.
450 * find the second name.
452 while ((second < line + llen) && isspace(*second))
453 second++;
455 if (line + llen <= second)
456 goto free_first_and_fail;
457 if (*second == '"') {
458 char *sp = unquote_c_style(second, NULL);
459 if (!sp)
460 goto free_first_and_fail;
461 cp = stop_at_slash(sp, strlen(sp));
462 if (!cp || cp == sp) {
463 free_both_and_fail:
464 free(sp);
465 goto free_first_and_fail;
467 /* They must match, otherwise ignore */
468 if (strcmp(cp+1, first))
469 goto free_both_and_fail;
470 free(sp);
471 return first;
474 /* unquoted second */
475 cp = stop_at_slash(second, line + llen - second);
476 if (!cp || cp == second)
477 goto free_first_and_fail;
478 cp++;
479 if (line + llen - cp != len + 1 ||
480 memcmp(first, cp, len))
481 goto free_first_and_fail;
482 return first;
485 /* unquoted first name */
486 name = stop_at_slash(line, llen);
487 if (!name || name == line)
488 return NULL;
490 name++;
492 /* since the first name is unquoted, a dq if exists must be
493 * the beginning of the second name.
495 for (second = name; second < line + llen; second++) {
496 if (*second == '"') {
497 const char *cp = second;
498 const char *np;
499 char *sp = unquote_c_style(second, NULL);
501 if (!sp)
502 return NULL;
503 np = stop_at_slash(sp, strlen(sp));
504 if (!np || np == sp) {
505 free_second_and_fail:
506 free(sp);
507 return NULL;
509 np++;
510 len = strlen(np);
511 if (len < cp - name &&
512 !strncmp(np, name, len) &&
513 isspace(name[len])) {
514 /* Good */
515 memmove(sp, np, len + 1);
516 return sp;
518 goto free_second_and_fail;
523 * Accept a name only if it shows up twice, exactly the same
524 * form.
526 for (len = 0 ; ; len++) {
527 char c = name[len];
529 switch (c) {
530 default:
531 continue;
532 case '\n':
533 return NULL;
534 case '\t': case ' ':
535 second = name+len;
536 for (;;) {
537 char c = *second++;
538 if (c == '\n')
539 return NULL;
540 if (c == '/')
541 break;
543 if (second[len] == '\n' && !memcmp(name, second, len)) {
544 char *ret = xmalloc(len + 1);
545 memcpy(ret, name, len);
546 ret[len] = 0;
547 return ret;
551 return NULL;
554 /* Verify that we recognize the lines following a git header */
555 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
557 unsigned long offset;
559 /* A git diff has explicit new/delete information, so we don't guess */
560 patch->is_new = 0;
561 patch->is_delete = 0;
564 * Some things may not have the old name in the
565 * rest of the headers anywhere (pure mode changes,
566 * or removing or adding empty files), so we get
567 * the default name from the header.
569 patch->def_name = git_header_name(line, len);
571 line += len;
572 size -= len;
573 linenr++;
574 for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
575 static const struct opentry {
576 const char *str;
577 int (*fn)(const char *, struct patch *);
578 } optable[] = {
579 { "@@ -", gitdiff_hdrend },
580 { "--- ", gitdiff_oldname },
581 { "+++ ", gitdiff_newname },
582 { "old mode ", gitdiff_oldmode },
583 { "new mode ", gitdiff_newmode },
584 { "deleted file mode ", gitdiff_delete },
585 { "new file mode ", gitdiff_newfile },
586 { "copy from ", gitdiff_copysrc },
587 { "copy to ", gitdiff_copydst },
588 { "rename old ", gitdiff_renamesrc },
589 { "rename new ", gitdiff_renamedst },
590 { "rename from ", gitdiff_renamesrc },
591 { "rename to ", gitdiff_renamedst },
592 { "similarity index ", gitdiff_similarity },
593 { "dissimilarity index ", gitdiff_dissimilarity },
594 { "index ", gitdiff_index },
595 { "", gitdiff_unrecognized },
597 int i;
599 len = linelen(line, size);
600 if (!len || line[len-1] != '\n')
601 break;
602 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
603 const struct opentry *p = optable + i;
604 int oplen = strlen(p->str);
605 if (len < oplen || memcmp(p->str, line, oplen))
606 continue;
607 if (p->fn(line + oplen, patch) < 0)
608 return offset;
609 break;
613 return offset;
616 static int parse_num(const char *line, unsigned long *p)
618 char *ptr;
620 if (!isdigit(*line))
621 return 0;
622 *p = strtoul(line, &ptr, 10);
623 return ptr - line;
626 static int parse_range(const char *line, int len, int offset, const char *expect,
627 unsigned long *p1, unsigned long *p2)
629 int digits, ex;
631 if (offset < 0 || offset >= len)
632 return -1;
633 line += offset;
634 len -= offset;
636 digits = parse_num(line, p1);
637 if (!digits)
638 return -1;
640 offset += digits;
641 line += digits;
642 len -= digits;
644 *p2 = *p1;
645 if (*line == ',') {
646 digits = parse_num(line+1, p2);
647 if (!digits)
648 return -1;
650 offset += digits+1;
651 line += digits+1;
652 len -= digits+1;
655 ex = strlen(expect);
656 if (ex > len)
657 return -1;
658 if (memcmp(line, expect, ex))
659 return -1;
661 return offset + ex;
665 * Parse a unified diff fragment header of the
666 * form "@@ -a,b +c,d @@"
668 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
670 int offset;
672 if (!len || line[len-1] != '\n')
673 return -1;
675 /* Figure out the number of lines in a fragment */
676 offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
677 offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
679 return offset;
682 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
684 unsigned long offset, len;
686 patch->is_rename = patch->is_copy = 0;
687 patch->is_new = patch->is_delete = -1;
688 patch->old_mode = patch->new_mode = 0;
689 patch->old_name = patch->new_name = NULL;
690 for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
691 unsigned long nextlen;
693 len = linelen(line, size);
694 if (!len)
695 break;
697 /* Testing this early allows us to take a few shortcuts.. */
698 if (len < 6)
699 continue;
702 * Make sure we don't find any unconnected patch fragmants.
703 * That's a sign that we didn't find a header, and that a
704 * patch has become corrupted/broken up.
706 if (!memcmp("@@ -", line, 4)) {
707 struct fragment dummy;
708 if (parse_fragment_header(line, len, &dummy) < 0)
709 continue;
710 error("patch fragment without header at line %d: %.*s", linenr, (int)len-1, line);
713 if (size < len + 6)
714 break;
717 * Git patch? It might not have a real patch, just a rename
718 * or mode change, so we handle that specially
720 if (!memcmp("diff --git ", line, 11)) {
721 int git_hdr_len = parse_git_header(line, len, size, patch);
722 if (git_hdr_len <= len)
723 continue;
724 if (!patch->old_name && !patch->new_name) {
725 if (!patch->def_name)
726 die("git diff header lacks filename information (line %d)", linenr);
727 patch->old_name = patch->new_name = patch->def_name;
729 *hdrsize = git_hdr_len;
730 return offset;
733 /** --- followed by +++ ? */
734 if (memcmp("--- ", line, 4) || memcmp("+++ ", line + len, 4))
735 continue;
738 * We only accept unified patches, so we want it to
739 * at least have "@@ -a,b +c,d @@\n", which is 14 chars
740 * minimum
742 nextlen = linelen(line + len, size - len);
743 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
744 continue;
746 /* Ok, we'll consider it a patch */
747 parse_traditional_patch(line, line+len, patch);
748 *hdrsize = len + nextlen;
749 linenr += 2;
750 return offset;
752 return -1;
756 * Parse a unified diff. Note that this really needs
757 * to parse each fragment separately, since the only
758 * way to know the difference between a "---" that is
759 * part of a patch, and a "---" that starts the next
760 * patch is to look at the line counts..
762 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
764 int added, deleted;
765 int len = linelen(line, size), offset;
766 unsigned long oldlines, newlines;
768 offset = parse_fragment_header(line, len, fragment);
769 if (offset < 0)
770 return -1;
771 oldlines = fragment->oldlines;
772 newlines = fragment->newlines;
774 if (patch->is_new < 0) {
775 patch->is_new = !oldlines;
776 if (!oldlines)
777 patch->old_name = NULL;
779 if (patch->is_delete < 0) {
780 patch->is_delete = !newlines;
781 if (!newlines)
782 patch->new_name = NULL;
785 if (patch->is_new != !oldlines)
786 return error("new file depends on old contents");
787 if (patch->is_delete != !newlines) {
788 if (newlines)
789 return error("deleted file still has contents");
790 fprintf(stderr, "** warning: file %s becomes empty but is not deleted\n", patch->new_name);
793 /* Parse the thing.. */
794 line += len;
795 size -= len;
796 linenr++;
797 added = deleted = 0;
798 for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
799 if (!oldlines && !newlines)
800 break;
801 len = linelen(line, size);
802 if (!len || line[len-1] != '\n')
803 return -1;
804 switch (*line) {
805 default:
806 return -1;
807 case ' ':
808 oldlines--;
809 newlines--;
810 break;
811 case '-':
812 deleted++;
813 oldlines--;
814 break;
815 case '+':
816 added++;
817 newlines--;
818 break;
820 /* We allow "\ No newline at end of file". Depending
821 * on locale settings when the patch was produced we
822 * don't know what this line looks like. The only
823 * thing we do know is that it begins with "\ ".
824 * Checking for 12 is just for sanity check -- any
825 * l10n of "\ No newline..." is at least that long.
827 case '\\':
828 if (len < 12 || memcmp(line, "\\ ", 2))
829 return -1;
830 break;
833 /* If a fragment ends with an incomplete line, we failed to include
834 * it in the above loop because we hit oldlines == newlines == 0
835 * before seeing it.
837 if (12 < size && !memcmp(line, "\\ ", 2))
838 offset += linelen(line, size);
840 patch->lines_added += added;
841 patch->lines_deleted += deleted;
842 return offset;
845 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
847 unsigned long offset = 0;
848 struct fragment **fragp = &patch->fragments;
850 while (size > 4 && !memcmp(line, "@@ -", 4)) {
851 struct fragment *fragment;
852 int len;
854 fragment = xmalloc(sizeof(*fragment));
855 memset(fragment, 0, sizeof(*fragment));
856 len = parse_fragment(line, size, patch, fragment);
857 if (len <= 0)
858 die("corrupt patch at line %d", linenr);
860 fragment->patch = line;
861 fragment->size = len;
863 *fragp = fragment;
864 fragp = &fragment->next;
866 offset += len;
867 line += len;
868 size -= len;
870 return offset;
873 static inline int metadata_changes(struct patch *patch)
875 return patch->is_rename > 0 ||
876 patch->is_copy > 0 ||
877 patch->is_new > 0 ||
878 patch->is_delete ||
879 (patch->old_mode && patch->new_mode &&
880 patch->old_mode != patch->new_mode);
883 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
885 int hdrsize, patchsize;
886 int offset = find_header(buffer, size, &hdrsize, patch);
888 if (offset < 0)
889 return offset;
891 patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
893 if (!patchsize && !metadata_changes(patch))
894 die("patch with only garbage at line %d", linenr);
896 return offset + hdrsize + patchsize;
899 static const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
900 static const char minuses[]= "----------------------------------------------------------------------";
902 static void show_stats(struct patch *patch)
904 const char *prefix = "";
905 char *name = patch->new_name;
906 char *qname = NULL;
907 int len, max, add, del, total;
909 if (!name)
910 name = patch->old_name;
912 if (0 < (len = quote_c_style(name, NULL, NULL, 0))) {
913 qname = xmalloc(len + 1);
914 quote_c_style(name, qname, NULL, 0);
915 name = qname;
919 * "scale" the filename
921 len = strlen(name);
922 max = max_len;
923 if (max > 50)
924 max = 50;
925 if (len > max) {
926 char *slash;
927 prefix = "...";
928 max -= 3;
929 name += len - max;
930 slash = strchr(name, '/');
931 if (slash)
932 name = slash;
934 len = max;
937 * scale the add/delete
939 max = max_change;
940 if (max + len > 70)
941 max = 70 - len;
943 add = patch->lines_added;
944 del = patch->lines_deleted;
945 total = add + del;
947 if (max_change > 0) {
948 total = (total * max + max_change / 2) / max_change;
949 add = (add * max + max_change / 2) / max_change;
950 del = total - add;
952 printf(" %s%-*s |%5d %.*s%.*s\n", prefix,
953 len, name, patch->lines_added + patch->lines_deleted,
954 add, pluses, del, minuses);
955 if (qname)
956 free(qname);
959 static int read_old_data(struct stat *st, const char *path, void *buf, unsigned long size)
961 int fd;
962 unsigned long got;
964 switch (st->st_mode & S_IFMT) {
965 case S_IFLNK:
966 return readlink(path, buf, size);
967 case S_IFREG:
968 fd = open(path, O_RDONLY);
969 if (fd < 0)
970 return error("unable to open %s", path);
971 got = 0;
972 for (;;) {
973 int ret = read(fd, buf + got, size - got);
974 if (ret < 0) {
975 if (errno == EAGAIN)
976 continue;
977 break;
979 if (!ret)
980 break;
981 got += ret;
983 close(fd);
984 return got;
986 default:
987 return -1;
991 static int find_offset(const char *buf, unsigned long size, const char *fragment, unsigned long fragsize, int line)
993 int i;
994 unsigned long start, backwards, forwards;
996 if (fragsize > size)
997 return -1;
999 start = 0;
1000 if (line > 1) {
1001 unsigned long offset = 0;
1002 i = line-1;
1003 while (offset + fragsize <= size) {
1004 if (buf[offset++] == '\n') {
1005 start = offset;
1006 if (!--i)
1007 break;
1012 /* Exact line number? */
1013 if (!memcmp(buf + start, fragment, fragsize))
1014 return start;
1017 * There's probably some smart way to do this, but I'll leave
1018 * that to the smart and beautiful people. I'm simple and stupid.
1020 backwards = start;
1021 forwards = start;
1022 for (i = 0; ; i++) {
1023 unsigned long try;
1024 int n;
1026 /* "backward" */
1027 if (i & 1) {
1028 if (!backwards) {
1029 if (forwards + fragsize > size)
1030 break;
1031 continue;
1033 do {
1034 --backwards;
1035 } while (backwards && buf[backwards-1] != '\n');
1036 try = backwards;
1037 } else {
1038 while (forwards + fragsize <= size) {
1039 if (buf[forwards++] == '\n')
1040 break;
1042 try = forwards;
1045 if (try + fragsize > size)
1046 continue;
1047 if (memcmp(buf + try, fragment, fragsize))
1048 continue;
1049 n = (i >> 1)+1;
1050 if (i & 1)
1051 n = -n;
1052 return try;
1056 * We should start searching forward and backward.
1058 return -1;
1061 struct buffer_desc {
1062 char *buffer;
1063 unsigned long size;
1064 unsigned long alloc;
1067 static int apply_one_fragment(struct buffer_desc *desc, struct fragment *frag)
1069 char *buf = desc->buffer;
1070 const char *patch = frag->patch;
1071 int offset, size = frag->size;
1072 char *old = xmalloc(size);
1073 char *new = xmalloc(size);
1074 int oldsize = 0, newsize = 0;
1076 while (size > 0) {
1077 int len = linelen(patch, size);
1078 int plen;
1080 if (!len)
1081 break;
1084 * "plen" is how much of the line we should use for
1085 * the actual patch data. Normally we just remove the
1086 * first character on the line, but if the line is
1087 * followed by "\ No newline", then we also remove the
1088 * last one (which is the newline, of course).
1090 plen = len-1;
1091 if (len < size && patch[len] == '\\')
1092 plen--;
1093 switch (*patch) {
1094 case ' ':
1095 case '-':
1096 memcpy(old + oldsize, patch + 1, plen);
1097 oldsize += plen;
1098 if (*patch == '-')
1099 break;
1100 /* Fall-through for ' ' */
1101 case '+':
1102 memcpy(new + newsize, patch + 1, plen);
1103 newsize += plen;
1104 break;
1105 case '@': case '\\':
1106 /* Ignore it, we already handled it */
1107 break;
1108 default:
1109 return -1;
1111 patch += len;
1112 size -= len;
1115 offset = find_offset(buf, desc->size, old, oldsize, frag->newpos);
1116 if (offset >= 0) {
1117 int diff = newsize - oldsize;
1118 unsigned long size = desc->size + diff;
1119 unsigned long alloc = desc->alloc;
1121 if (size > alloc) {
1122 alloc = size + 8192;
1123 desc->alloc = alloc;
1124 buf = xrealloc(buf, alloc);
1125 desc->buffer = buf;
1127 desc->size = size;
1128 memmove(buf + offset + newsize, buf + offset + oldsize, size - offset - newsize);
1129 memcpy(buf + offset, new, newsize);
1130 offset = 0;
1133 free(old);
1134 free(new);
1135 return offset;
1138 static int apply_fragments(struct buffer_desc *desc, struct patch *patch)
1140 struct fragment *frag = patch->fragments;
1142 while (frag) {
1143 if (apply_one_fragment(desc, frag) < 0)
1144 return error("patch failed: %s:%ld", patch->old_name, frag->oldpos);
1145 frag = frag->next;
1147 return 0;
1150 static int apply_data(struct patch *patch, struct stat *st)
1152 char *buf;
1153 unsigned long size, alloc;
1154 struct buffer_desc desc;
1156 size = 0;
1157 alloc = 0;
1158 buf = NULL;
1159 if (patch->old_name) {
1160 size = st->st_size;
1161 alloc = size + 8192;
1162 buf = xmalloc(alloc);
1163 if (read_old_data(st, patch->old_name, buf, alloc) != size)
1164 return error("read of %s failed", patch->old_name);
1167 desc.size = size;
1168 desc.alloc = alloc;
1169 desc.buffer = buf;
1170 if (apply_fragments(&desc, patch) < 0)
1171 return -1;
1172 patch->result = desc.buffer;
1173 patch->resultsize = desc.size;
1175 if (patch->is_delete && patch->resultsize)
1176 return error("removal patch leaves file contents");
1178 return 0;
1181 static int check_patch(struct patch *patch)
1183 struct stat st;
1184 const char *old_name = patch->old_name;
1185 const char *new_name = patch->new_name;
1187 if (old_name) {
1188 int changed;
1189 int stat_ret = lstat(old_name, &st);
1191 if (check_index) {
1192 int pos = cache_name_pos(old_name, strlen(old_name));
1193 if (pos < 0)
1194 return error("%s: does not exist in index",
1195 old_name);
1196 if (stat_ret < 0) {
1197 struct checkout costate;
1198 if (errno != ENOENT)
1199 return error("%s: %s", old_name,
1200 strerror(errno));
1201 /* checkout */
1202 costate.base_dir = "";
1203 costate.base_dir_len = 0;
1204 costate.force = 0;
1205 costate.quiet = 0;
1206 costate.not_new = 0;
1207 costate.refresh_cache = 1;
1208 if (checkout_entry(active_cache[pos],
1209 &costate) ||
1210 lstat(old_name, &st))
1211 return -1;
1214 changed = ce_match_stat(active_cache[pos], &st);
1215 if (changed)
1216 return error("%s: does not match index",
1217 old_name);
1219 else if (stat_ret < 0)
1220 return error("%s: %s", old_name, strerror(errno));
1222 if (patch->is_new < 0)
1223 patch->is_new = 0;
1224 st.st_mode = ntohl(create_ce_mode(st.st_mode));
1225 if (!patch->old_mode)
1226 patch->old_mode = st.st_mode;
1227 if ((st.st_mode ^ patch->old_mode) & S_IFMT)
1228 return error("%s: wrong type", old_name);
1229 if (st.st_mode != patch->old_mode)
1230 fprintf(stderr, "warning: %s has type %o, expected %o\n",
1231 old_name, st.st_mode, patch->old_mode);
1234 if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
1235 if (check_index && cache_name_pos(new_name, strlen(new_name)) >= 0)
1236 return error("%s: already exists in index", new_name);
1237 if (!lstat(new_name, &st))
1238 return error("%s: already exists in working directory", new_name);
1239 if (errno != ENOENT)
1240 return error("%s: %s", new_name, strerror(errno));
1241 if (!patch->new_mode) {
1242 if (patch->is_new)
1243 patch->new_mode = S_IFREG | 0644;
1244 else
1245 patch->new_mode = patch->old_mode;
1249 if (new_name && old_name) {
1250 int same = !strcmp(old_name, new_name);
1251 if (!patch->new_mode)
1252 patch->new_mode = patch->old_mode;
1253 if ((patch->old_mode ^ patch->new_mode) & S_IFMT)
1254 return error("new mode (%o) of %s does not match old mode (%o)%s%s",
1255 patch->new_mode, new_name, patch->old_mode,
1256 same ? "" : " of ", same ? "" : old_name);
1259 if (apply_data(patch, &st) < 0)
1260 return error("%s: patch does not apply", old_name);
1261 return 0;
1264 static int check_patch_list(struct patch *patch)
1266 int error = 0;
1268 for (;patch ; patch = patch->next)
1269 error |= check_patch(patch);
1270 return error;
1273 static inline int is_null_sha1(const unsigned char *sha1)
1275 return !memcmp(sha1, null_sha1, 20);
1278 static void show_index_list(struct patch *list)
1280 struct patch *patch;
1282 /* Once we start supporting the reverse patch, it may be
1283 * worth showing the new sha1 prefix, but until then...
1285 for (patch = list; patch; patch = patch->next) {
1286 const unsigned char *sha1_ptr;
1287 unsigned char sha1[20];
1288 const char *name;
1290 name = patch->old_name ? patch->old_name : patch->new_name;
1291 if (patch->is_new)
1292 sha1_ptr = null_sha1;
1293 else if (get_sha1(patch->old_sha1_prefix, sha1))
1294 die("sha1 information is lacking or useless (%s).",
1295 name);
1296 else
1297 sha1_ptr = sha1;
1299 printf("%06o %s ",patch->old_mode, sha1_to_hex(sha1_ptr));
1300 if (line_termination && quote_c_style(name, NULL, NULL, 0))
1301 quote_c_style(name, NULL, stdout, 0);
1302 else
1303 fputs(name, stdout);
1304 putchar(line_termination);
1308 static void stat_patch_list(struct patch *patch)
1310 int files, adds, dels;
1312 for (files = adds = dels = 0 ; patch ; patch = patch->next) {
1313 files++;
1314 adds += patch->lines_added;
1315 dels += patch->lines_deleted;
1316 show_stats(patch);
1319 printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
1322 static void numstat_patch_list(struct patch *patch)
1324 for ( ; patch; patch = patch->next) {
1325 const char *name;
1326 name = patch->old_name ? patch->old_name : patch->new_name;
1327 printf("%d\t%d\t", patch->lines_added, patch->lines_deleted);
1328 if (line_termination && quote_c_style(name, NULL, NULL, 0))
1329 quote_c_style(name, NULL, stdout, 0);
1330 else
1331 fputs(name, stdout);
1332 putchar('\n');
1336 static void show_file_mode_name(const char *newdelete, unsigned int mode, const char *name)
1338 if (mode)
1339 printf(" %s mode %06o %s\n", newdelete, mode, name);
1340 else
1341 printf(" %s %s\n", newdelete, name);
1344 static void show_mode_change(struct patch *p, int show_name)
1346 if (p->old_mode && p->new_mode && p->old_mode != p->new_mode) {
1347 if (show_name)
1348 printf(" mode change %06o => %06o %s\n",
1349 p->old_mode, p->new_mode, p->new_name);
1350 else
1351 printf(" mode change %06o => %06o\n",
1352 p->old_mode, p->new_mode);
1356 static void show_rename_copy(struct patch *p)
1358 const char *renamecopy = p->is_rename ? "rename" : "copy";
1359 const char *old, *new;
1361 /* Find common prefix */
1362 old = p->old_name;
1363 new = p->new_name;
1364 while (1) {
1365 const char *slash_old, *slash_new;
1366 slash_old = strchr(old, '/');
1367 slash_new = strchr(new, '/');
1368 if (!slash_old ||
1369 !slash_new ||
1370 slash_old - old != slash_new - new ||
1371 memcmp(old, new, slash_new - new))
1372 break;
1373 old = slash_old + 1;
1374 new = slash_new + 1;
1376 /* p->old_name thru old is the common prefix, and old and new
1377 * through the end of names are renames
1379 if (old != p->old_name)
1380 printf(" %s %.*s{%s => %s} (%d%%)\n", renamecopy,
1381 (int)(old - p->old_name), p->old_name,
1382 old, new, p->score);
1383 else
1384 printf(" %s %s => %s (%d%%)\n", renamecopy,
1385 p->old_name, p->new_name, p->score);
1386 show_mode_change(p, 0);
1389 static void summary_patch_list(struct patch *patch)
1391 struct patch *p;
1393 for (p = patch; p; p = p->next) {
1394 if (p->is_new)
1395 show_file_mode_name("create", p->new_mode, p->new_name);
1396 else if (p->is_delete)
1397 show_file_mode_name("delete", p->old_mode, p->old_name);
1398 else {
1399 if (p->is_rename || p->is_copy)
1400 show_rename_copy(p);
1401 else {
1402 if (p->score) {
1403 printf(" rewrite %s (%d%%)\n",
1404 p->new_name, p->score);
1405 show_mode_change(p, 0);
1407 else
1408 show_mode_change(p, 1);
1414 static void patch_stats(struct patch *patch)
1416 int lines = patch->lines_added + patch->lines_deleted;
1418 if (lines > max_change)
1419 max_change = lines;
1420 if (patch->old_name) {
1421 int len = quote_c_style(patch->old_name, NULL, NULL, 0);
1422 if (!len)
1423 len = strlen(patch->old_name);
1424 if (len > max_len)
1425 max_len = len;
1427 if (patch->new_name) {
1428 int len = quote_c_style(patch->new_name, NULL, NULL, 0);
1429 if (!len)
1430 len = strlen(patch->new_name);
1431 if (len > max_len)
1432 max_len = len;
1436 static void remove_file(struct patch *patch)
1438 if (write_index) {
1439 if (remove_file_from_cache(patch->old_name) < 0)
1440 die("unable to remove %s from index", patch->old_name);
1442 unlink(patch->old_name);
1445 static void add_index_file(const char *path, unsigned mode, void *buf, unsigned long size)
1447 struct stat st;
1448 struct cache_entry *ce;
1449 int namelen = strlen(path);
1450 unsigned ce_size = cache_entry_size(namelen);
1452 if (!write_index)
1453 return;
1455 ce = xmalloc(ce_size);
1456 memset(ce, 0, ce_size);
1457 memcpy(ce->name, path, namelen);
1458 ce->ce_mode = create_ce_mode(mode);
1459 ce->ce_flags = htons(namelen);
1460 if (lstat(path, &st) < 0)
1461 die("unable to stat newly created file %s", path);
1462 fill_stat_cache_info(ce, &st);
1463 if (write_sha1_file(buf, size, "blob", ce->sha1) < 0)
1464 die("unable to create backing store for newly created file %s", path);
1465 if (add_cache_entry(ce, ADD_CACHE_OK_TO_ADD) < 0)
1466 die("unable to add cache entry for %s", path);
1469 static void create_subdirectories(const char *path)
1471 int len = strlen(path);
1472 char *buf = xmalloc(len + 1);
1473 const char *slash = path;
1475 while ((slash = strchr(slash+1, '/')) != NULL) {
1476 len = slash - path;
1477 memcpy(buf, path, len);
1478 buf[len] = 0;
1479 if (mkdir(buf, 0777) < 0) {
1480 if (errno != EEXIST)
1481 break;
1484 free(buf);
1487 static int try_create_file(const char *path, unsigned int mode, const char *buf, unsigned long size)
1489 int fd;
1491 if (S_ISLNK(mode))
1492 return symlink(buf, path);
1493 fd = open(path, O_CREAT | O_EXCL | O_WRONLY | O_TRUNC, (mode & 0100) ? 0777 : 0666);
1494 if (fd < 0)
1495 return -1;
1496 while (size) {
1497 int written = write(fd, buf, size);
1498 if (written < 0) {
1499 if (errno == EINTR || errno == EAGAIN)
1500 continue;
1501 die("writing file %s: %s", path, strerror(errno));
1503 if (!written)
1504 die("out of space writing file %s", path);
1505 buf += written;
1506 size -= written;
1508 if (close(fd) < 0)
1509 die("closing file %s: %s", path, strerror(errno));
1510 return 0;
1514 * We optimistically assume that the directories exist,
1515 * which is true 99% of the time anyway. If they don't,
1516 * we create them and try again.
1518 static void create_one_file(const char *path, unsigned mode, const char *buf, unsigned long size)
1520 if (!try_create_file(path, mode, buf, size))
1521 return;
1523 if (errno == ENOENT) {
1524 create_subdirectories(path);
1525 if (!try_create_file(path, mode, buf, size))
1526 return;
1529 if (errno == EEXIST) {
1530 unsigned int nr = getpid();
1532 for (;;) {
1533 const char *newpath;
1534 newpath = mkpath("%s~%u", path, nr);
1535 if (!try_create_file(newpath, mode, buf, size)) {
1536 if (!rename(newpath, path))
1537 return;
1538 unlink(newpath);
1539 break;
1541 if (errno != EEXIST)
1542 break;
1545 die("unable to write file %s mode %o", path, mode);
1548 static void create_file(struct patch *patch)
1550 const char *path = patch->new_name;
1551 unsigned mode = patch->new_mode;
1552 unsigned long size = patch->resultsize;
1553 char *buf = patch->result;
1555 if (!mode)
1556 mode = S_IFREG | 0644;
1557 create_one_file(path, mode, buf, size);
1558 add_index_file(path, mode, buf, size);
1561 static void write_out_one_result(struct patch *patch)
1563 if (patch->is_delete > 0) {
1564 remove_file(patch);
1565 return;
1567 if (patch->is_new > 0 || patch->is_copy) {
1568 create_file(patch);
1569 return;
1572 * Rename or modification boils down to the same
1573 * thing: remove the old, write the new
1575 remove_file(patch);
1576 create_file(patch);
1579 static void write_out_results(struct patch *list, int skipped_patch)
1581 if (!list && !skipped_patch)
1582 die("No changes");
1584 while (list) {
1585 write_out_one_result(list);
1586 list = list->next;
1590 static struct cache_file cache_file;
1592 static struct excludes {
1593 struct excludes *next;
1594 const char *path;
1595 } *excludes;
1597 static int use_patch(struct patch *p)
1599 const char *pathname = p->new_name ? p->new_name : p->old_name;
1600 struct excludes *x = excludes;
1601 while (x) {
1602 if (fnmatch(x->path, pathname, 0) == 0)
1603 return 0;
1604 x = x->next;
1606 return 1;
1609 static int apply_patch(int fd)
1611 int newfd;
1612 unsigned long offset, size;
1613 char *buffer = read_patch_file(fd, &size);
1614 struct patch *list = NULL, **listp = &list;
1615 int skipped_patch = 0;
1617 if (!buffer)
1618 return -1;
1619 offset = 0;
1620 while (size > 0) {
1621 struct patch *patch;
1622 int nr;
1624 patch = xmalloc(sizeof(*patch));
1625 memset(patch, 0, sizeof(*patch));
1626 nr = parse_chunk(buffer + offset, size, patch);
1627 if (nr < 0)
1628 break;
1629 if (use_patch(patch)) {
1630 patch_stats(patch);
1631 *listp = patch;
1632 listp = &patch->next;
1633 } else {
1634 /* perhaps free it a bit better? */
1635 free(patch);
1636 skipped_patch++;
1638 offset += nr;
1639 size -= nr;
1642 newfd = -1;
1643 write_index = check_index && apply;
1644 if (write_index)
1645 newfd = hold_index_file_for_update(&cache_file, get_index_file());
1646 if (check_index) {
1647 if (read_cache() < 0)
1648 die("unable to read index file");
1651 if ((check || apply) && check_patch_list(list) < 0)
1652 exit(1);
1654 if (apply)
1655 write_out_results(list, skipped_patch);
1657 if (write_index) {
1658 if (write_cache(newfd, active_cache, active_nr) ||
1659 commit_index_file(&cache_file))
1660 die("Unable to write new cachefile");
1663 if (show_index_info)
1664 show_index_list(list);
1666 if (diffstat)
1667 stat_patch_list(list);
1669 if (numstat)
1670 numstat_patch_list(list);
1672 if (summary)
1673 summary_patch_list(list);
1675 free(buffer);
1676 return 0;
1679 int main(int argc, char **argv)
1681 int i;
1682 int read_stdin = 1;
1684 for (i = 1; i < argc; i++) {
1685 const char *arg = argv[i];
1686 int fd;
1688 if (!strcmp(arg, "-")) {
1689 apply_patch(0);
1690 read_stdin = 0;
1691 continue;
1693 if (!strncmp(arg, "--exclude=", 10)) {
1694 struct excludes *x = xmalloc(sizeof(*x));
1695 x->path = arg + 10;
1696 x->next = excludes;
1697 excludes = x;
1698 continue;
1700 if (!strcmp(arg, "--stat")) {
1701 apply = 0;
1702 diffstat = 1;
1703 continue;
1705 if (!strcmp(arg, "--numstat")) {
1706 apply = 0;
1707 numstat = 1;
1708 continue;
1710 if (!strcmp(arg, "--summary")) {
1711 apply = 0;
1712 summary = 1;
1713 continue;
1715 if (!strcmp(arg, "--check")) {
1716 apply = 0;
1717 check = 1;
1718 continue;
1720 if (!strcmp(arg, "--index")) {
1721 check_index = 1;
1722 continue;
1724 if (!strcmp(arg, "--apply")) {
1725 apply = 1;
1726 continue;
1728 if (!strcmp(arg, "--index-info")) {
1729 apply = 0;
1730 show_index_info = 1;
1731 continue;
1733 if (!strcmp(arg, "-z")) {
1734 line_termination = 0;
1735 continue;
1737 fd = open(arg, O_RDONLY);
1738 if (fd < 0)
1739 usage(apply_usage);
1740 read_stdin = 0;
1741 apply_patch(fd);
1742 close(fd);
1744 if (read_stdin)
1745 apply_patch(0);
1746 return 0;