2 * This handles recursive filename detection with exclude
3 * files, index knowledge etc..
5 * Copyright (C) Linus Torvalds, 2005-2006
6 * Junio Hamano, 2005-2006
12 struct path_simplify {
17 static int read_directory_recursive(struct dir_struct *dir, const char *path, int len,
18 int check_only, const struct path_simplify *simplify);
19 static int get_dtype(struct dirent *de, const char *path, int len);
21 static int common_prefix(const char **pathspec)
23 const char *path, *slash, *next;
30 slash = strrchr(path, '/');
34 prefix = slash - path + 1;
35 while ((next = *++pathspec) != NULL) {
36 int len = strlen(next);
37 if (len >= prefix && !memcmp(path, next, prefix))
43 if (next[--len] != '/')
45 if (memcmp(path, next, len+1))
54 int fill_directory(struct dir_struct *dir, const char **pathspec)
60 * Calculate common prefix for the pathspec, and
61 * use that to optimize the directory walk
63 len = common_prefix(pathspec);
67 path = xmemdupz(*pathspec, len);
69 /* Read the directory and prune it */
70 read_directory(dir, path, len, pathspec);
75 * Does 'match' match the given name?
78 * (1) the 'match' string is leading directory of 'name', or
79 * (2) the 'match' string is a wildcard and matches 'name', or
80 * (3) the 'match' string is exactly the same as 'name'.
82 * and the return value tells which case it was.
84 * It returns 0 when there is no match.
86 static int match_one(const char *match, const char *name, int namelen)
90 /* If the match was just the prefix, we matched */
92 return MATCHED_RECURSIVELY;
95 unsigned char c1 = *match;
96 unsigned char c2 = *name;
97 if (c1 == '\0' || is_glob_special(c1))
108 * If we don't match the matchstring exactly,
109 * we need to match by fnmatch
111 matchlen = strlen(match);
112 if (strncmp(match, name, matchlen))
113 return !fnmatch(match, name, 0) ? MATCHED_FNMATCH : 0;
115 if (namelen == matchlen)
116 return MATCHED_EXACTLY;
117 if (match[matchlen-1] == '/' || name[matchlen] == '/')
118 return MATCHED_RECURSIVELY;
123 * Given a name and a list of pathspecs, see if the name matches
124 * any of the pathspecs. The caller is also interested in seeing
125 * all pathspec matches some names it calls this function with
126 * (otherwise the user could have mistyped the unmatched pathspec),
127 * and a mark is left in seen[] array for pathspec element that
128 * actually matched anything.
130 int match_pathspec(const char **pathspec, const char *name, int namelen,
131 int prefix, char *seen)
141 for (i = 0; pathspec[i] != NULL; i++) {
143 const char *match = pathspec[i] + prefix;
144 if (seen && seen[i] == MATCHED_EXACTLY)
146 how = match_one(match, name, namelen);
150 if (seen && seen[i] < how)
157 static int no_wildcard(const char *string)
159 return string[strcspn(string, "*?[{\\")] == '\0';
162 void add_exclude(const char *string, const char *base,
163 int baselen, struct exclude_list *which)
170 if (*string == '!') {
174 len = strlen(string);
175 if (len && string[len - 1] == '/') {
177 x = xmalloc(sizeof(*x) + len);
179 memcpy(s, string, len - 1);
183 flags = EXC_FLAG_MUSTBEDIR;
185 x = xmalloc(sizeof(*x));
188 x->to_exclude = to_exclude;
189 x->patternlen = strlen(string);
191 x->baselen = baselen;
193 if (!strchr(string, '/'))
194 x->flags |= EXC_FLAG_NODIR;
195 if (no_wildcard(string))
196 x->flags |= EXC_FLAG_NOWILDCARD;
197 if (*string == '*' && no_wildcard(string+1))
198 x->flags |= EXC_FLAG_ENDSWITH;
199 ALLOC_GROW(which->excludes, which->nr + 1, which->alloc);
200 which->excludes[which->nr++] = x;
203 static int add_excludes_from_file_1(const char *fname,
207 struct exclude_list *which)
214 fd = open(fname, O_RDONLY);
215 if (fd < 0 || fstat(fd, &st) < 0)
217 size = xsize_t(st.st_size);
222 buf = xmalloc(size+1);
223 if (read_in_full(fd, buf, size) != size)
234 for (i = 0; i < size; i++) {
235 if (buf[i] == '\n') {
236 if (entry != buf + i && entry[0] != '#') {
237 buf[i - (i && buf[i-1] == '\r')] = 0;
238 add_exclude(entry, base, baselen, which);
251 void add_excludes_from_file(struct dir_struct *dir, const char *fname)
253 if (add_excludes_from_file_1(fname, "", 0, NULL,
254 &dir->exclude_list[EXC_FILE]) < 0)
255 die("cannot use %s as an exclude file", fname);
258 static void prep_exclude(struct dir_struct *dir, const char *base, int baselen)
260 struct exclude_list *el;
261 struct exclude_stack *stk = NULL;
264 if ((!dir->exclude_per_dir) ||
265 (baselen + strlen(dir->exclude_per_dir) >= PATH_MAX))
266 return; /* too long a path -- ignore */
268 /* Pop the ones that are not the prefix of the path being checked. */
269 el = &dir->exclude_list[EXC_DIRS];
270 while ((stk = dir->exclude_stack) != NULL) {
271 if (stk->baselen <= baselen &&
272 !strncmp(dir->basebuf, base, stk->baselen))
274 dir->exclude_stack = stk->prev;
275 while (stk->exclude_ix < el->nr)
276 free(el->excludes[--el->nr]);
281 /* Read from the parent directories and push them down. */
282 current = stk ? stk->baselen : -1;
283 while (current < baselen) {
284 struct exclude_stack *stk = xcalloc(1, sizeof(*stk));
292 cp = strchr(base + current + 1, '/');
294 die("oops in prep_exclude");
297 stk->prev = dir->exclude_stack;
298 stk->baselen = cp - base;
299 stk->exclude_ix = el->nr;
300 memcpy(dir->basebuf + current, base + current,
301 stk->baselen - current);
302 strcpy(dir->basebuf + stk->baselen, dir->exclude_per_dir);
303 add_excludes_from_file_1(dir->basebuf,
304 dir->basebuf, stk->baselen,
306 dir->exclude_stack = stk;
307 current = stk->baselen;
309 dir->basebuf[baselen] = '\0';
312 /* Scan the list and let the last match determine the fate.
313 * Return 1 for exclude, 0 for include and -1 for undecided.
315 static int excluded_1(const char *pathname,
316 int pathlen, const char *basename, int *dtype,
317 struct exclude_list *el)
322 for (i = el->nr - 1; 0 <= i; i--) {
323 struct exclude *x = el->excludes[i];
324 const char *exclude = x->pattern;
325 int to_exclude = x->to_exclude;
327 if (x->flags & EXC_FLAG_MUSTBEDIR) {
328 if (*dtype == DT_UNKNOWN)
329 *dtype = get_dtype(NULL, pathname, pathlen);
330 if (*dtype != DT_DIR)
334 if (x->flags & EXC_FLAG_NODIR) {
336 if (x->flags & EXC_FLAG_NOWILDCARD) {
337 if (!strcmp(exclude, basename))
339 } else if (x->flags & EXC_FLAG_ENDSWITH) {
340 if (x->patternlen - 1 <= pathlen &&
341 !strcmp(exclude + 1, pathname + pathlen - x->patternlen + 1))
344 if (fnmatch(exclude, basename, 0) == 0)
349 /* match with FNM_PATHNAME:
350 * exclude has base (baselen long) implicitly
353 int baselen = x->baselen;
357 if (pathlen < baselen ||
358 (baselen && pathname[baselen-1] != '/') ||
359 strncmp(pathname, x->base, baselen))
362 if (x->flags & EXC_FLAG_NOWILDCARD) {
363 if (!strcmp(exclude, pathname + baselen))
366 if (fnmatch(exclude, pathname+baselen,
373 return -1; /* undecided */
376 int excluded(struct dir_struct *dir, const char *pathname, int *dtype_p)
378 int pathlen = strlen(pathname);
380 const char *basename = strrchr(pathname, '/');
381 basename = (basename) ? basename+1 : pathname;
383 prep_exclude(dir, pathname, basename-pathname);
384 for (st = EXC_CMDL; st <= EXC_FILE; st++) {
385 switch (excluded_1(pathname, pathlen, basename,
386 dtype_p, &dir->exclude_list[st])) {
396 static struct dir_entry *dir_entry_new(const char *pathname, int len)
398 struct dir_entry *ent;
400 ent = xmalloc(sizeof(*ent) + len + 1);
402 memcpy(ent->name, pathname, len);
407 static struct dir_entry *dir_add_name(struct dir_struct *dir, const char *pathname, int len)
409 if (cache_name_exists(pathname, len, ignore_case))
412 ALLOC_GROW(dir->entries, dir->nr+1, dir->alloc);
413 return dir->entries[dir->nr++] = dir_entry_new(pathname, len);
416 static struct dir_entry *dir_add_ignored(struct dir_struct *dir, const char *pathname, int len)
418 if (!cache_name_is_other(pathname, len))
421 ALLOC_GROW(dir->ignored, dir->ignored_nr+1, dir->ignored_alloc);
422 return dir->ignored[dir->ignored_nr++] = dir_entry_new(pathname, len);
426 index_nonexistent = 0,
432 * The index sorts alphabetically by entry name, which
433 * means that a gitlink sorts as '\0' at the end, while
434 * a directory (which is defined not as an entry, but as
435 * the files it contains) will sort with the '/' at the
438 static enum exist_status directory_exists_in_index(const char *dirname, int len)
440 int pos = cache_name_pos(dirname, len);
443 while (pos < active_nr) {
444 struct cache_entry *ce = active_cache[pos++];
445 unsigned char endchar;
447 if (strncmp(ce->name, dirname, len))
449 endchar = ce->name[len];
453 return index_directory;
454 if (!endchar && S_ISGITLINK(ce->ce_mode))
457 return index_nonexistent;
461 * When we find a directory when traversing the filesystem, we
462 * have three distinct cases:
465 * - see it as a directory
468 * and which one we choose depends on a combination of existing
469 * git index contents and the flags passed into the directory
472 * Case 1: If we *already* have entries in the index under that
473 * directory name, we always recurse into the directory to see
476 * Case 2: If we *already* have that directory name as a gitlink,
477 * we always continue to see it as a gitlink, regardless of whether
478 * there is an actual git directory there or not (it might not
479 * be checked out as a subproject!)
481 * Case 3: if we didn't have it in the index previously, we
482 * have a few sub-cases:
484 * (a) if "show_other_directories" is true, we show it as
485 * just a directory, unless "hide_empty_directories" is
486 * also true and the directory is empty, in which case
487 * we just ignore it entirely.
488 * (b) if it looks like a git directory, and we don't have
489 * 'no_gitlinks' set we treat it as a gitlink, and show it
491 * (c) otherwise, we recurse into it.
493 enum directory_treatment {
496 recurse_into_directory,
499 static enum directory_treatment treat_directory(struct dir_struct *dir,
500 const char *dirname, int len,
501 const struct path_simplify *simplify)
503 /* The "len-1" is to strip the final '/' */
504 switch (directory_exists_in_index(dirname, len-1)) {
505 case index_directory:
506 return recurse_into_directory;
509 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
510 return ignore_directory;
511 return show_directory;
513 case index_nonexistent:
514 if (dir->flags & DIR_SHOW_OTHER_DIRECTORIES)
516 if (!(dir->flags & DIR_NO_GITLINKS)) {
517 unsigned char sha1[20];
518 if (resolve_gitlink_ref(dirname, "HEAD", sha1) == 0)
519 return show_directory;
521 return recurse_into_directory;
524 /* This is the "show_other_directories" case */
525 if (!(dir->flags & DIR_HIDE_EMPTY_DIRECTORIES))
526 return show_directory;
527 if (!read_directory_recursive(dir, dirname, len, 1, simplify))
528 return ignore_directory;
529 return show_directory;
533 * This is an inexact early pruning of any recursive directory
534 * reading - if the path cannot possibly be in the pathspec,
535 * return true, and we'll skip it early.
537 static int simplify_away(const char *path, int pathlen, const struct path_simplify *simplify)
541 const char *match = simplify->path;
542 int len = simplify->len;
548 if (!memcmp(path, match, len))
557 static int in_pathspec(const char *path, int len, const struct path_simplify *simplify)
560 for (; simplify->path; simplify++) {
561 if (len == simplify->len
562 && !memcmp(path, simplify->path, len))
569 static int get_index_dtype(const char *path, int len)
572 struct cache_entry *ce;
574 ce = cache_name_exists(path, len, 0);
576 if (!ce_uptodate(ce))
578 if (S_ISGITLINK(ce->ce_mode))
581 * Nobody actually cares about the
582 * difference between DT_LNK and DT_REG
587 /* Try to look it up as a directory */
588 pos = cache_name_pos(path, len);
592 while (pos < active_nr) {
593 ce = active_cache[pos++];
594 if (strncmp(ce->name, path, len))
596 if (ce->name[len] > '/')
598 if (ce->name[len] < '/')
600 if (!ce_uptodate(ce))
601 break; /* continue? */
607 static int get_dtype(struct dirent *de, const char *path, int len)
609 int dtype = de ? DTYPE(de) : DT_UNKNOWN;
612 if (dtype != DT_UNKNOWN)
614 dtype = get_index_dtype(path, len);
615 if (dtype != DT_UNKNOWN)
617 if (lstat(path, &st))
619 if (S_ISREG(st.st_mode))
621 if (S_ISDIR(st.st_mode))
623 if (S_ISLNK(st.st_mode))
629 * Read a directory tree. We currently ignore anything but
630 * directories, regular files and symlinks. That's because git
631 * doesn't handle them at all yet. Maybe that will change some
634 * Also, we ignore the name ".git" (even if it is not a directory).
635 * That likely will not change.
637 static int read_directory_recursive(struct dir_struct *dir, const char *base, int baselen, int check_only, const struct path_simplify *simplify)
639 DIR *fdir = opendir(*base ? base : ".");
644 char path[PATH_MAX + 1];
645 memcpy(path, base, baselen);
647 while ((de = readdir(fdir)) != NULL) {
651 if (is_dot_or_dotdot(de->d_name) ||
652 !strcmp(de->d_name, ".git"))
654 len = strlen(de->d_name);
655 /* Ignore overly long pathnames! */
656 if (len + baselen + 8 > sizeof(path))
658 memcpy(path + baselen, de->d_name, len+1);
660 if (simplify_away(path, len, simplify))
664 exclude = excluded(dir, path, &dtype);
665 if (exclude && (dir->flags & DIR_COLLECT_IGNORED)
666 && in_pathspec(path, len, simplify))
667 dir_add_ignored(dir, path,len);
670 * Excluded? If we don't explicitly want to show
671 * ignored files, ignore it
673 if (exclude && !(dir->flags & DIR_SHOW_IGNORED))
676 if (dtype == DT_UNKNOWN)
677 dtype = get_dtype(de, path, len);
680 * Do we want to see just the ignored files?
681 * We still need to recurse into directories,
682 * even if we don't ignore them, since the
683 * directory may contain files that we do..
685 if (!exclude && (dir->flags & DIR_SHOW_IGNORED)) {
694 memcpy(path + len, "/", 2);
696 switch (treat_directory(dir, path, len, simplify)) {
698 if (exclude != !!(dir->flags
702 case recurse_into_directory:
703 contents += read_directory_recursive(dir,
704 path, len, 0, simplify);
706 case ignore_directory:
718 dir_add_name(dir, path, len);
727 static int cmp_name(const void *p1, const void *p2)
729 const struct dir_entry *e1 = *(const struct dir_entry **)p1;
730 const struct dir_entry *e2 = *(const struct dir_entry **)p2;
732 return cache_name_compare(e1->name, e1->len,
737 * Return the length of the "simple" part of a path match limiter.
739 static int simple_length(const char *match)
744 unsigned char c = *match++;
746 if (c == '\0' || is_glob_special(c))
751 static struct path_simplify *create_simplify(const char **pathspec)
754 struct path_simplify *simplify = NULL;
759 for (nr = 0 ; ; nr++) {
762 alloc = alloc_nr(alloc);
763 simplify = xrealloc(simplify, alloc * sizeof(*simplify));
768 simplify[nr].path = match;
769 simplify[nr].len = simple_length(match);
771 simplify[nr].path = NULL;
772 simplify[nr].len = 0;
776 static void free_simplify(struct path_simplify *simplify)
781 int read_directory(struct dir_struct *dir, const char *path, int len, const char **pathspec)
783 struct path_simplify *simplify;
785 if (has_symlink_leading_path(path, len))
788 simplify = create_simplify(pathspec);
789 read_directory_recursive(dir, path, len, 0, simplify);
790 free_simplify(simplify);
791 qsort(dir->entries, dir->nr, sizeof(struct dir_entry *), cmp_name);
792 qsort(dir->ignored, dir->ignored_nr, sizeof(struct dir_entry *), cmp_name);
796 int file_exists(const char *f)
799 return lstat(f, &sb) == 0;
803 * get_relative_cwd() gets the prefix of the current working directory
804 * relative to 'dir'. If we are not inside 'dir', it returns NULL.
806 * As a convenience, it also returns NULL if 'dir' is already NULL. The
807 * reason for this behaviour is that it is natural for functions returning
808 * directory names to return NULL to say "this directory does not exist"
809 * or "this directory is invalid". These cases are usually handled the
810 * same as if the cwd is not inside 'dir' at all, so get_relative_cwd()
811 * returns NULL for both of them.
813 * Most notably, get_relative_cwd(buffer, size, get_git_work_tree())
814 * unifies the handling of "outside work tree" with "no work tree at all".
816 char *get_relative_cwd(char *buffer, int size, const char *dir)
822 if (!getcwd(buffer, size))
823 die_errno("can't find the current directory");
825 if (!is_absolute_path(dir))
826 dir = make_absolute_path(dir);
828 while (*dir && *dir == *cwd) {
839 int is_inside_dir(const char *dir)
841 char buffer[PATH_MAX];
842 return get_relative_cwd(buffer, sizeof(buffer), dir) != NULL;
845 int is_empty_dir(const char *path)
847 DIR *dir = opendir(path);
854 while ((e = readdir(dir)) != NULL)
855 if (!is_dot_or_dotdot(e->d_name)) {
864 int remove_dir_recursively(struct strbuf *path, int flag)
868 int ret = 0, original_len = path->len, len;
869 int only_empty = (flag & REMOVE_DIR_EMPTY_ONLY);
870 unsigned char submodule_head[20];
872 if ((flag & REMOVE_DIR_KEEP_NESTED_GIT) &&
873 !resolve_gitlink_ref(path->buf, "HEAD", submodule_head))
874 /* Do not descend and nuke a nested git work tree. */
877 dir = opendir(path->buf);
880 if (path->buf[original_len - 1] != '/')
881 strbuf_addch(path, '/');
884 while ((e = readdir(dir)) != NULL) {
886 if (is_dot_or_dotdot(e->d_name))
889 strbuf_setlen(path, len);
890 strbuf_addstr(path, e->d_name);
891 if (lstat(path->buf, &st))
893 else if (S_ISDIR(st.st_mode)) {
894 if (!remove_dir_recursively(path, only_empty))
895 continue; /* happy */
896 } else if (!only_empty && !unlink(path->buf))
897 continue; /* happy, too */
899 /* path too long, stat fails, or non-directory still exists */
905 strbuf_setlen(path, original_len);
907 ret = rmdir(path->buf);
911 void setup_standard_excludes(struct dir_struct *dir)
915 dir->exclude_per_dir = ".gitignore";
916 path = git_path("info/exclude");
917 if (!access(path, R_OK))
918 add_excludes_from_file(dir, path);
919 if (excludes_file && !access(excludes_file, R_OK))
920 add_excludes_from_file(dir, excludes_file);
923 int remove_path(const char *name)
927 if (unlink(name) && errno != ENOENT)
930 slash = strrchr(name, '/');
932 char *dirs = xstrdup(name);
933 slash = dirs + (slash - name);
936 } while (rmdir(dirs) && (slash = strrchr(dirs, '/')));