[PATCH] mkdelta enhancements (take 2)
[git/mingw.git] / apply.c
blob3160207b927ce76d7039e2a71f958784a4938eb7
1 /*
2 * apply.c
4 * Copyright (C) Linus Torvalds, 2005
6 * This applies patches on top of some (arbitrary) version of the SCM.
8 * NOTE! It does all its work in the index file, and only cares about
9 * the files in the working directory if you tell it to "merge" the
10 * patch apply.
12 * Even when merging it always takes the source from the index, and
13 * uses the working tree as a "branch" for a 3-way merge.
15 #include <ctype.h>
17 #include "cache.h"
19 // We default to the merge behaviour, since that's what most people would
20 // expect.
22 // --check turns on checking that the working tree matches the
23 // files that are being modified, but doesn't apply the patch
24 // --stat does just a diffstat, and doesn't actually apply
25 // --show-files shows the directory changes
27 static int merge_patch = 1;
28 static int diffstat = 0;
29 static int check = 0;
30 static int apply = 1;
31 static int show_files = 0;
32 static const char apply_usage[] = "git-apply [--stat] [--check] [--show-files] <patch>";
35 * For "diff-stat" like behaviour, we keep track of the biggest change
36 * we've seen, and the longest filename. That allows us to do simple
37 * scaling.
39 static int max_change, max_len;
42 * Various "current state", notably line numbers and what
43 * file (and how) we're patching right now.. The "is_xxxx"
44 * things are flags, where -1 means "don't know yet".
46 static int linenr = 1;
48 struct fragment {
49 unsigned long oldpos, oldlines;
50 unsigned long newpos, newlines;
51 const char *patch;
52 int size;
53 struct fragment *next;
56 struct patch {
57 char *new_name, *old_name, *def_name;
58 unsigned int old_mode, new_mode;
59 int is_rename, is_copy, is_new, is_delete;
60 int lines_added, lines_deleted;
61 struct fragment *fragments;
62 struct patch *next;
65 #define CHUNKSIZE (8192)
66 #define SLOP (16)
68 static void *read_patch_file(int fd, unsigned long *sizep)
70 unsigned long size = 0, alloc = CHUNKSIZE;
71 void *buffer = xmalloc(alloc);
73 for (;;) {
74 int nr = alloc - size;
75 if (nr < 1024) {
76 alloc += CHUNKSIZE;
77 buffer = xrealloc(buffer, alloc);
78 nr = alloc - size;
80 nr = read(fd, buffer + size, nr);
81 if (!nr)
82 break;
83 if (nr < 0) {
84 if (errno == EAGAIN)
85 continue;
86 die("git-apply: read returned %s", strerror(errno));
88 size += nr;
90 *sizep = size;
93 * Make sure that we have some slop in the buffer
94 * so that we can do speculative "memcmp" etc, and
95 * see to it that it is NUL-filled.
97 if (alloc < size + SLOP)
98 buffer = xrealloc(buffer, size + SLOP);
99 memset(buffer + size, 0, SLOP);
100 return buffer;
103 static unsigned long linelen(char *buffer, unsigned long size)
105 unsigned long len = 0;
106 while (size--) {
107 len++;
108 if (*buffer++ == '\n')
109 break;
111 return len;
114 static int is_dev_null(const char *str)
116 return !memcmp("/dev/null", str, 9) && isspace(str[9]);
119 #define TERM_EXIST 1
120 #define TERM_SPACE 2
121 #define TERM_TAB 4
123 static int name_terminate(const char *name, int namelen, int c, int terminate)
125 if (c == ' ' && !(terminate & TERM_SPACE))
126 return 0;
127 if (c == '\t' && !(terminate & TERM_TAB))
128 return 0;
131 * Do we want an existing name? Return false and
132 * continue if it's not there.
134 if (terminate & TERM_EXIST)
135 return cache_name_pos(name, namelen) >= 0;
137 return 1;
140 static char * find_name(const char *line, char *def, int p_value, int terminate)
142 int len;
143 const char *start = line;
144 char *name;
146 for (;;) {
147 char c = *line;
149 if (isspace(c)) {
150 if (c == '\n')
151 break;
152 if (name_terminate(start, line-start, c, terminate))
153 break;
155 line++;
156 if (c == '/' && !--p_value)
157 start = line;
159 if (!start)
160 return def;
161 len = line - start;
162 if (!len)
163 return def;
166 * Generally we prefer the shorter name, especially
167 * if the other one is just a variation of that with
168 * something else tacked on to the end (ie "file.orig"
169 * or "file~").
171 if (def) {
172 int deflen = strlen(def);
173 if (deflen < len && !strncmp(start, def, deflen))
174 return def;
177 name = xmalloc(len + 1);
178 memcpy(name, start, len);
179 name[len] = 0;
180 free(def);
181 return name;
185 * Get the name etc info from the --/+++ lines of a traditional patch header
187 * NOTE! This hardcodes "-p1" behaviour in filename detection.
189 * FIXME! The end-of-filename heuristics are kind of screwy. For existing
190 * files, we can happily check the index for a match, but for creating a
191 * new file we should try to match whatever "patch" does. I have no idea.
193 static void parse_traditional_patch(const char *first, const char *second, struct patch *patch)
195 int p_value = 1;
196 char *name;
198 first += 4; // skip "--- "
199 second += 4; // skip "+++ "
200 if (is_dev_null(first)) {
201 patch->is_new = 1;
202 patch->is_delete = 0;
203 name = find_name(second, NULL, p_value, TERM_SPACE | TERM_TAB);
204 patch->new_name = name;
205 } else if (is_dev_null(second)) {
206 patch->is_new = 0;
207 patch->is_delete = 1;
208 name = find_name(first, NULL, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
209 patch->old_name = name;
210 } else {
211 name = find_name(first, NULL, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
212 name = find_name(second, name, p_value, TERM_EXIST | TERM_SPACE | TERM_TAB);
213 patch->old_name = patch->new_name = name;
215 if (!name)
216 die("unable to find filename in patch at line %d", linenr);
219 static int gitdiff_hdrend(const char *line, struct patch *patch)
221 return -1;
225 * We're anal about diff header consistency, to make
226 * sure that we don't end up having strange ambiguous
227 * patches floating around.
229 * As a result, gitdiff_{old|new}name() will check
230 * their names against any previous information, just
231 * to make sure..
233 static char *gitdiff_verify_name(const char *line, int isnull, char *orig_name, const char *oldnew)
235 int len;
236 const char *name;
238 if (!orig_name && !isnull)
239 return find_name(line, NULL, 1, 0);
241 name = "/dev/null";
242 len = 9;
243 if (orig_name) {
244 name = orig_name;
245 len = strlen(name);
246 if (isnull)
247 die("git-apply: bad git-diff - expected /dev/null, got %s on line %d", name, linenr);
250 if (*name == '/')
251 goto absolute_path;
253 for (;;) {
254 char c = *line++;
255 if (c == '\n')
256 break;
257 if (c != '/')
258 continue;
259 absolute_path:
260 if (memcmp(line, name, len) || line[len] != '\n')
261 break;
262 return orig_name;
264 die("git-apply: bad git-diff - inconsistent %s filename on line %d", oldnew, linenr);
265 return NULL;
268 static int gitdiff_oldname(const char *line, struct patch *patch)
270 patch->old_name = gitdiff_verify_name(line, patch->is_new, patch->old_name, "old");
271 return 0;
274 static int gitdiff_newname(const char *line, struct patch *patch)
276 patch->new_name = gitdiff_verify_name(line, patch->is_delete, patch->new_name, "new");
277 return 0;
280 static int gitdiff_oldmode(const char *line, struct patch *patch)
282 patch->old_mode = strtoul(line, NULL, 8);
283 return 0;
286 static int gitdiff_newmode(const char *line, struct patch *patch)
288 patch->new_mode = strtoul(line, NULL, 8);
289 return 0;
292 static int gitdiff_delete(const char *line, struct patch *patch)
294 patch->is_delete = 1;
295 patch->old_name = patch->def_name;
296 return gitdiff_oldmode(line, patch);
299 static int gitdiff_newfile(const char *line, struct patch *patch)
301 patch->is_new = 1;
302 patch->new_name = patch->def_name;
303 return gitdiff_newmode(line, patch);
306 static int gitdiff_copysrc(const char *line, struct patch *patch)
308 patch->is_copy = 1;
309 patch->old_name = find_name(line, NULL, 0, 0);
310 return 0;
313 static int gitdiff_copydst(const char *line, struct patch *patch)
315 patch->is_copy = 1;
316 patch->new_name = find_name(line, NULL, 0, 0);
317 return 0;
320 static int gitdiff_renamesrc(const char *line, struct patch *patch)
322 patch->is_rename = 1;
323 patch->old_name = find_name(line, NULL, 0, 0);
324 return 0;
327 static int gitdiff_renamedst(const char *line, struct patch *patch)
329 patch->is_rename = 1;
330 patch->new_name = find_name(line, NULL, 0, 0);
331 return 0;
334 static int gitdiff_similarity(const char *line, struct patch *patch)
336 return 0;
340 * This is normal for a diff that doesn't change anything: we'll fall through
341 * into the next diff. Tell the parser to break out.
343 static int gitdiff_unrecognized(const char *line, struct patch *patch)
345 return -1;
348 static char *git_header_name(char *line)
350 int len;
351 char *name, *second;
354 * Find the first '/'
356 name = line;
357 for (;;) {
358 char c = *name++;
359 if (c == '\n')
360 return NULL;
361 if (c == '/')
362 break;
366 * We don't accept absolute paths (/dev/null) as possibly valid
368 if (name == line+1)
369 return NULL;
372 * Accept a name only if it shows up twice, exactly the same
373 * form.
375 for (len = 0 ; ; len++) {
376 char c = name[len];
378 switch (c) {
379 default:
380 continue;
381 case '\n':
382 break;
383 case '\t': case ' ':
384 second = name+len;
385 for (;;) {
386 char c = *second++;
387 if (c == '\n')
388 return NULL;
389 if (c == '/')
390 break;
392 if (second[len] == '\n' && !memcmp(name, second, len)) {
393 char *ret = xmalloc(len + 1);
394 memcpy(ret, name, len);
395 ret[len] = 0;
396 return ret;
400 return NULL;
403 /* Verify that we recognize the lines following a git header */
404 static int parse_git_header(char *line, int len, unsigned int size, struct patch *patch)
406 unsigned long offset;
408 /* A git diff has explicit new/delete information, so we don't guess */
409 patch->is_new = 0;
410 patch->is_delete = 0;
413 * Some things may not have the old name in the
414 * rest of the headers anywhere (pure mode changes,
415 * or removing or adding empty files), so we get
416 * the default name from the header.
418 patch->def_name = git_header_name(line + strlen("diff --git "));
420 line += len;
421 size -= len;
422 linenr++;
423 for (offset = len ; size > 0 ; offset += len, size -= len, line += len, linenr++) {
424 static const struct opentry {
425 const char *str;
426 int (*fn)(const char *, struct patch *);
427 } optable[] = {
428 { "@@ -", gitdiff_hdrend },
429 { "--- ", gitdiff_oldname },
430 { "+++ ", gitdiff_newname },
431 { "old mode ", gitdiff_oldmode },
432 { "new mode ", gitdiff_newmode },
433 { "deleted file mode ", gitdiff_delete },
434 { "new file mode ", gitdiff_newfile },
435 { "copy from ", gitdiff_copysrc },
436 { "copy to ", gitdiff_copydst },
437 { "rename from ", gitdiff_renamesrc },
438 { "rename to ", gitdiff_renamedst },
439 { "similarity index ", gitdiff_similarity },
440 { "", gitdiff_unrecognized },
442 int i;
444 len = linelen(line, size);
445 if (!len || line[len-1] != '\n')
446 break;
447 for (i = 0; i < sizeof(optable) / sizeof(optable[0]); i++) {
448 const struct opentry *p = optable + i;
449 int oplen = strlen(p->str);
450 if (len < oplen || memcmp(p->str, line, oplen))
451 continue;
452 if (p->fn(line + oplen, patch) < 0)
453 return offset;
454 break;
458 return offset;
461 static int parse_num(const char *line, unsigned long *p)
463 char *ptr;
465 if (!isdigit(*line))
466 return 0;
467 *p = strtoul(line, &ptr, 10);
468 return ptr - line;
471 static int parse_range(const char *line, int len, int offset, const char *expect,
472 unsigned long *p1, unsigned long *p2)
474 int digits, ex;
476 if (offset < 0 || offset >= len)
477 return -1;
478 line += offset;
479 len -= offset;
481 digits = parse_num(line, p1);
482 if (!digits)
483 return -1;
485 offset += digits;
486 line += digits;
487 len -= digits;
489 *p2 = *p1;
490 if (*line == ',') {
491 digits = parse_num(line+1, p2);
492 if (!digits)
493 return -1;
495 offset += digits+1;
496 line += digits+1;
497 len -= digits+1;
500 ex = strlen(expect);
501 if (ex > len)
502 return -1;
503 if (memcmp(line, expect, ex))
504 return -1;
506 return offset + ex;
510 * Parse a unified diff fragment header of the
511 * form "@@ -a,b +c,d @@"
513 static int parse_fragment_header(char *line, int len, struct fragment *fragment)
515 int offset;
517 if (!len || line[len-1] != '\n')
518 return -1;
520 /* Figure out the number of lines in a fragment */
521 offset = parse_range(line, len, 4, " +", &fragment->oldpos, &fragment->oldlines);
522 offset = parse_range(line, len, offset, " @@", &fragment->newpos, &fragment->newlines);
524 return offset;
527 static int find_header(char *line, unsigned long size, int *hdrsize, struct patch *patch)
529 unsigned long offset, len;
531 patch->is_rename = patch->is_copy = 0;
532 patch->is_new = patch->is_delete = -1;
533 patch->old_mode = patch->new_mode = 0;
534 patch->old_name = patch->new_name = NULL;
535 for (offset = 0; size > 0; offset += len, size -= len, line += len, linenr++) {
536 unsigned long nextlen;
538 len = linelen(line, size);
539 if (!len)
540 break;
542 /* Testing this early allows us to take a few shortcuts.. */
543 if (len < 6)
544 continue;
547 * Make sure we don't find any unconnected patch fragmants.
548 * That's a sign that we didn't find a header, and that a
549 * patch has become corrupted/broken up.
551 if (!memcmp("@@ -", line, 4)) {
552 struct fragment dummy;
553 if (parse_fragment_header(line, len, &dummy) < 0)
554 continue;
555 error("patch fragment without header at line %d: %.*s", linenr, len-1, line);
558 if (size < len + 6)
559 break;
562 * Git patch? It might not have a real patch, just a rename
563 * or mode change, so we handle that specially
565 if (!memcmp("diff --git ", line, 11)) {
566 int git_hdr_len = parse_git_header(line, len, size, patch);
567 if (git_hdr_len < 0)
568 continue;
569 if (!patch->old_name && !patch->new_name)
570 die("git diff header lacks filename information");
571 *hdrsize = git_hdr_len;
572 return offset;
575 /** --- followed by +++ ? */
576 if (memcmp("--- ", line, 4) || memcmp("+++ ", line + len, 4))
577 continue;
580 * We only accept unified patches, so we want it to
581 * at least have "@@ -a,b +c,d @@\n", which is 14 chars
582 * minimum
584 nextlen = linelen(line + len, size - len);
585 if (size < nextlen + 14 || memcmp("@@ -", line + len + nextlen, 4))
586 continue;
588 /* Ok, we'll consider it a patch */
589 parse_traditional_patch(line, line+len, patch);
590 *hdrsize = len + nextlen;
591 linenr += 2;
592 return offset;
594 return -1;
598 * Parse a unified diff. Note that this really needs
599 * to parse each fragment separately, since the only
600 * way to know the difference between a "---" that is
601 * part of a patch, and a "---" that starts the next
602 * patch is to look at the line counts..
604 static int parse_fragment(char *line, unsigned long size, struct patch *patch, struct fragment *fragment)
606 int added, deleted;
607 int len = linelen(line, size), offset;
608 unsigned long pos[4], oldlines, newlines;
610 offset = parse_fragment_header(line, len, fragment);
611 if (offset < 0)
612 return -1;
613 oldlines = fragment->oldlines;
614 newlines = fragment->newlines;
616 if (patch->is_new < 0 && (pos[0] || oldlines))
617 patch->is_new = 0;
618 if (patch->is_delete < 0 && (pos[1] || newlines))
619 patch->is_delete = 0;
621 /* Parse the thing.. */
622 line += len;
623 size -= len;
624 linenr++;
625 added = deleted = 0;
626 for (offset = len; size > 0; offset += len, size -= len, line += len, linenr++) {
627 if (!oldlines && !newlines)
628 break;
629 len = linelen(line, size);
630 if (!len || line[len-1] != '\n')
631 return -1;
632 switch (*line) {
633 default:
634 return -1;
635 case ' ':
636 oldlines--;
637 newlines--;
638 break;
639 case '-':
640 deleted++;
641 oldlines--;
642 break;
643 case '+':
644 added++;
645 newlines--;
646 break;
647 /* We allow "\ No newline at end of file" */
648 case '\\':
649 break;
652 patch->lines_added += added;
653 patch->lines_deleted += deleted;
654 return offset;
657 static int parse_single_patch(char *line, unsigned long size, struct patch *patch)
659 unsigned long offset = 0;
660 struct fragment **fragp = &patch->fragments;
662 while (size > 4 && !memcmp(line, "@@ -", 4)) {
663 struct fragment *fragment;
664 int len;
666 fragment = xmalloc(sizeof(*fragment));
667 memset(fragment, 0, sizeof(*fragment));
668 len = parse_fragment(line, size, patch, fragment);
669 if (len <= 0)
670 die("corrupt patch at line %d", linenr);
672 fragment->patch = line;
673 fragment->size = len;
675 *fragp = fragment;
676 fragp = &fragment->next;
678 offset += len;
679 line += len;
680 size -= len;
682 return offset;
685 static int parse_chunk(char *buffer, unsigned long size, struct patch *patch)
687 int hdrsize, patchsize;
688 int offset = find_header(buffer, size, &hdrsize, patch);
690 if (offset < 0)
691 return offset;
693 patchsize = parse_single_patch(buffer + offset + hdrsize, size - offset - hdrsize, patch);
695 return offset + hdrsize + patchsize;
698 const char pluses[] = "++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++";
699 const char minuses[]= "----------------------------------------------------------------------";
701 static void show_stats(struct patch *patch)
703 char *name = patch->old_name;
704 int len, max, add, del;
706 if (!name)
707 name = patch->new_name;
710 * "scale" the filename
712 len = strlen(name);
713 max = max_len;
714 if (max > 50)
715 max = 50;
716 if (len > max)
717 name += len - max;
718 len = max;
721 * scale the add/delete
723 max = max_change;
724 if (max + len > 70)
725 max = 70 - len;
727 add = (patch->lines_added * max + max_change/2) / max_change;
728 del = (patch->lines_deleted * max + max_change/2) / max_change;
729 printf(" %-*s |%5d %.*s%.*s\n",
730 len, name, patch->lines_added + patch->lines_deleted,
731 add, pluses, del, minuses);
734 static int check_patch(struct patch *patch)
736 struct stat st;
737 const char *old_name = patch->old_name;
738 const char *new_name = patch->new_name;
740 if (old_name) {
741 int pos = cache_name_pos(old_name, strlen(old_name));
742 int changed;
744 if (pos < 0)
745 return error("%s: does not exist in index", old_name);
746 if (patch->is_new < 0)
747 patch->is_new = 0;
748 if (lstat(old_name, &st) < 0)
749 return error("%s: %s\n", strerror(errno));
750 changed = ce_match_stat(active_cache[pos], &st);
751 if (changed)
752 return error("%s: does not match index", old_name);
753 if (!patch->old_mode)
754 patch->old_mode = st.st_mode;
757 if (new_name && (patch->is_new | patch->is_rename | patch->is_copy)) {
758 if (cache_name_pos(new_name, strlen(new_name)) >= 0)
759 return error("%s: already exists in index", new_name);
760 if (!lstat(new_name, &st))
761 return error("%s: already exists in working directory", new_name);
762 if (errno != ENOENT)
763 return error("%s: %s", new_name, strerror(errno));
765 return 0;
768 static int check_patch_list(struct patch *patch)
770 int error = 0;
772 for (;patch ; patch = patch->next)
773 error |= check_patch(patch);
774 return error;
777 static void show_file(int c, unsigned int mode, const char *name)
779 printf("%c %o %s\n", c, mode, name);
782 static void show_file_list(struct patch *patch)
784 for (;patch ; patch = patch->next) {
785 if (patch->is_rename) {
786 show_file('-', patch->old_mode, patch->old_name);
787 show_file('+', patch->new_mode, patch->new_name);
788 continue;
790 if (patch->is_copy || patch->is_new) {
791 show_file('+', patch->new_mode, patch->new_name);
792 continue;
794 if (patch->is_delete) {
795 show_file('-', patch->old_mode, patch->old_name);
796 continue;
798 if (patch->old_mode && patch->new_mode && patch->old_mode != patch->new_mode) {
799 printf("M %o:%o %s\n", patch->old_mode, patch->new_mode, patch->old_name);
800 continue;
802 printf("M %o %s\n", patch->old_mode, patch->old_name);
806 static void stat_patch_list(struct patch *patch)
808 int files, adds, dels;
810 for (files = adds = dels = 0 ; patch ; patch = patch->next) {
811 files++;
812 adds += patch->lines_added;
813 dels += patch->lines_deleted;
814 show_stats(patch);
817 printf(" %d files changed, %d insertions(+), %d deletions(-)\n", files, adds, dels);
820 static void patch_stats(struct patch *patch)
822 int lines = patch->lines_added + patch->lines_deleted;
824 if (lines > max_change)
825 max_change = lines;
826 if (patch->old_name) {
827 int len = strlen(patch->old_name);
828 if (len > max_len)
829 max_len = len;
831 if (patch->new_name) {
832 int len = strlen(patch->new_name);
833 if (len > max_len)
834 max_len = len;
838 static int apply_patch(int fd)
840 unsigned long offset, size;
841 char *buffer = read_patch_file(fd, &size);
842 struct patch *list = NULL, **listp = &list;
844 if (!buffer)
845 return -1;
846 offset = 0;
847 while (size > 0) {
848 struct patch *patch;
849 int nr;
851 patch = xmalloc(sizeof(*patch));
852 memset(patch, 0, sizeof(*patch));
853 nr = parse_chunk(buffer + offset, size, patch);
854 if (nr < 0)
855 break;
856 patch_stats(patch);
857 *listp = patch;
858 listp = &patch->next;
859 offset += nr;
860 size -= nr;
863 if ((check || apply) && check_patch_list(list) < 0)
864 exit(1);
866 if (show_files)
867 show_file_list(list);
869 if (diffstat)
870 stat_patch_list(list);
872 free(buffer);
873 return 0;
876 int main(int argc, char **argv)
878 int i;
879 int read_stdin = 1;
881 if (read_cache() < 0)
882 die("unable to read index file");
884 for (i = 1; i < argc; i++) {
885 const char *arg = argv[i];
886 int fd;
888 if (!strcmp(arg, "-")) {
889 apply_patch(0);
890 read_stdin = 0;
891 continue;
893 if (!strcmp(arg, "--no-merge")) {
894 merge_patch = 0;
895 continue;
897 if (!strcmp(arg, "--stat")) {
898 apply = 0;
899 diffstat = 1;
900 continue;
902 if (!strcmp(arg, "--check")) {
903 apply = 0;
904 check = 1;
905 continue;
907 if (!strcmp(arg, "--show-files")) {
908 show_files = 1;
909 continue;
911 fd = open(arg, O_RDONLY);
912 if (fd < 0)
913 usage(apply_usage);
914 read_stdin = 0;
915 apply_patch(fd);
916 close(fd);
918 if (read_stdin)
919 apply_patch(0);
920 return 0;