diff options
author | Ulrich Drepper <drepper@redhat.com> | 2007-02-14 19:52:38 +0000 |
---|---|---|
committer | Ulrich Drepper <drepper@redhat.com> | 2007-02-14 19:52:38 +0000 |
commit | 05f135ba3b85e0bf2449df6a678f001453d975ce (patch) | |
tree | 5f531106f88278f165d2e58e7985ccf639775fc8 /posix | |
parent | 2f69996772e39953951a3c5d8df885c775601060 (diff) | |
download | glibc-05f135ba3b85e0bf2449df6a678f001453d975ce.zip glibc-05f135ba3b85e0bf2449df6a678f001453d975ce.tar.gz glibc-05f135ba3b85e0bf2449df6a678f001453d975ce.tar.bz2 |
[BZ #3996]
2007-02-14 Jakub Jelinek <jakub@redhat.com>
[BZ #3996]
* posix/glob.c (attribute_hidden): Define if not defined.
(glob): Unescape dirname, filename or username when needed and not
GLOB_NOESCAPE. Handle \/ correctly. Handle GLOB_MARK if filename
is NULL. Handle unescaped [ in pattern without closing ].
Don't pass GLOB_CHECK down to recursive glob for directories.
(__glob_pattern_type): New function.
(__glob_pattern_p): Implement using __glob_pattern_type.
(glob_in_dir): Handle GLOB_NOCHECK patterns containing no meta
characters and backslashes if not GLOB_NOESCAPE or unterminated [.
Remove unreachable code.
* posix/globtest.sh: Add a couple of new tests.
Diffstat (limited to 'posix')
-rw-r--r-- | posix/glob.c | 377 | ||||
-rwxr-xr-x | posix/globtest.sh | 71 |
2 files changed, 331 insertions, 117 deletions
diff --git a/posix/glob.c b/posix/glob.c index 0079a15..6d8a891 100644 --- a/posix/glob.c +++ b/posix/glob.c @@ -1,4 +1,5 @@ -/* Copyright (C) 1991-2002,2003,2004,2005,2006 Free Software Foundation, Inc. +/* Copyright (C) 1991-2002, 2003, 2004, 2005, 2006, 2007 + Free Software Foundation, Inc. This file is part of the GNU C Library. The GNU C Library is free software; you can redistribute it and/or @@ -192,9 +193,15 @@ static const char *next_brace_sub (const char *begin, int flags) __THROW; #endif /* !defined _LIBC || !defined GLOB_ONLY_P */ +#ifndef attribute_hidden +# define attribute_hidden +#endif + static int glob_in_dir (const char *pattern, const char *directory, int flags, int (*errfunc) (const char *, int), glob_t *pglob); +extern int __glob_pattern_type (const char *pattern, int quote) + attribute_hidden; #if !defined _LIBC || !defined GLOB_ONLY_P static int prefix_array (const char *prefix, char **array, size_t n) __THROW; @@ -250,6 +257,9 @@ glob (pattern, flags, errfunc, pglob) size_t dirlen; int status; size_t oldcount; + int meta; + int dirname_modified; + glob_t dirs; if (pattern == NULL || pglob == NULL || (flags & ~__GLOB_FLAGS) != 0) { @@ -418,6 +428,7 @@ glob (pattern, flags, errfunc, pglob) if (filename == NULL) filename = strchr (pattern, ':'); #endif /* __MSDOS__ || WINDOWS32 */ + dirname_modified = 0; if (filename == NULL) { /* This can mean two things: a simple name or "~name". The latter @@ -486,10 +497,32 @@ glob (pattern, flags, errfunc, pglob) && dirlen > 1) /* "pattern/". Expand "pattern", appending slashes. */ { + int orig_flags = flags; + if (!(flags & GLOB_NOESCAPE) && dirname[dirlen - 1] == '\\') + { + /* "pattern\\/". Remove the final backslash if it hasn't + been quoted. */ + char *p = (char *) &dirname[dirlen - 1]; + + while (p > dirname && p[-1] == '\\') --p; + if ((&dirname[dirlen] - p) & 1) + { + *(char *) &dirname[--dirlen] = '\0'; + flags &= ~(GLOB_NOCHECK | GLOB_NOMAGIC); + } + } int val = glob (dirname, flags | GLOB_MARK, errfunc, pglob); if (val == 0) pglob->gl_flags = ((pglob->gl_flags & ~GLOB_MARK) | (flags & GLOB_MARK)); + else if (val == GLOB_NOMATCH && flags != orig_flags) + { + /* Make sure globfree (&dirs); is a nop. */ + dirs.gl_pathv = NULL; + flags = orig_flags; + oldcount = pglob->gl_pathc + pglob->gl_offs; + goto no_matches; + } return val; } } @@ -517,7 +550,9 @@ glob (pattern, flags, errfunc, pglob) #ifndef VMS if ((flags & (GLOB_TILDE|GLOB_TILDE_CHECK)) && dirname[0] == '~') { - if (dirname[1] == '\0' || dirname[1] == '/') + if (dirname[1] == '\0' || dirname[1] == '/' + || (!(flags & GLOB_NOESCAPE) && dirname[1] == '\\' + && (dirname[2] == '\0' || dirname[2] == '/'))) { /* Look up home directory. */ const char *home_dir = getenv ("HOME"); @@ -594,7 +629,10 @@ glob (pattern, flags, errfunc, pglob) # endif /* Now construct the full directory. */ if (dirname[1] == '\0') - dirname = home_dir; + { + dirname = home_dir; + dirlen = strlen (dirname); + } else { char *newp; @@ -603,7 +641,9 @@ glob (pattern, flags, errfunc, pglob) mempcpy (mempcpy (newp, home_dir, home_len), &dirname[1], dirlen); dirname = newp; + dirlen += home_len - 1; } + dirname_modified = 1; } # if !defined _AMIGA && !defined WINDOWS32 else @@ -611,15 +651,52 @@ glob (pattern, flags, errfunc, pglob) char *end_name = strchr (dirname, '/'); const char *user_name; const char *home_dir; + char *unescape = NULL; + if (!(flags & GLOB_NOESCAPE)) + { + if (end_name == NULL) + { + unescape = strchr (dirname, '\\'); + if (unescape) + end_name = strchr (unescape, '\0'); + } + else + unescape = memchr (dirname, '\\', end_name - dirname); + } if (end_name == NULL) user_name = dirname + 1; else { char *newp; newp = (char *) __alloca (end_name - dirname); - *((char *) mempcpy (newp, dirname + 1, end_name - dirname)) - = '\0'; + if (unescape != NULL) + { + char *p = mempcpy (newp, dirname + 1, + unescape - dirname - 1); + char *q = unescape; + while (*q != '\0') + { + if (*q == '\\') + { + if (q[1] == '\0') + { + /* "~fo\\o\\" unescape to user_name "foo\\", + but "~fo\\o\\/" unescape to user_name + "foo". */ + if (filename == NULL) + *p++ = '\\'; + break; + } + ++q; + } + *p++ = *q++; + } + *p = '\0'; + } + else + *((char *) mempcpy (newp, dirname + 1, end_name - dirname)) + = '\0'; user_name = newp; } @@ -673,6 +750,8 @@ glob (pattern, flags, errfunc, pglob) *((char *) mempcpy (mempcpy (newp, home_dir, home_len), end_name, rest_len)) = '\0'; dirname = newp; + dirlen = home_len + rest_len; + dirname_modified = 1; } else if (flags & GLOB_TILDE_CHECK) @@ -714,9 +793,22 @@ glob (pattern, flags, errfunc, pglob) } pglob->gl_pathv = new_gl_pathv; - pglob->gl_pathv[newcount] = strdup (dirname); - if (pglob->gl_pathv[newcount] == NULL) - goto nospace; + if (flags & GLOB_MARK) + { + char *p; + pglob->gl_pathv[newcount] = malloc (dirlen + 2); + if (pglob->gl_pathv[newcount] == NULL) + goto nospace; + p = mempcpy (pglob->gl_pathv[newcount], dirname, dirlen); + p[0] = '/'; + p[1] = '\0'; + } + else + { + pglob->gl_pathv[newcount] = strdup (dirname); + if (pglob->gl_pathv[newcount] == NULL) + goto nospace; + } pglob->gl_pathv[++newcount] = NULL; ++pglob->gl_pathc; pglob->gl_flags = flags; @@ -728,14 +820,30 @@ glob (pattern, flags, errfunc, pglob) return GLOB_NOMATCH; } - if (__glob_pattern_p (dirname, !(flags & GLOB_NOESCAPE))) + meta = __glob_pattern_type (dirname, !(flags & GLOB_NOESCAPE)); + /* meta is 1 if correct glob pattern containing metacharacters. + If meta has bit (1 << 2) set, it means there was an unterminated + [ which we handle the same, using fnmatch. Broken unterminated + pattern bracket expressions ought to be rare enough that it is + not worth special casing them, fnmatch will do the right thing. */ + if (meta & 5) { /* The directory name contains metacharacters, so we have to glob for the directory, and then glob for the pattern in each directory found. */ - glob_t dirs; size_t i; + if (!(flags & GLOB_NOESCAPE) && dirlen > 0 && dirname[dirlen - 1] == '\\') + { + /* "foo\\/bar". Remove the final backslash from dirname + if it has not been quoted. */ + char *p = (char *) &dirname[dirlen - 1]; + + while (p > dirname && p[-1] == '\\') --p; + if ((&dirname[dirlen] - p) & 1) + *(char *) &dirname[--dirlen] = '\0'; + } + if ((flags & GLOB_ALTDIRFUNC) != 0) { /* Use the alternative access functions also in the recursive @@ -748,12 +856,16 @@ glob (pattern, flags, errfunc, pglob) } status = glob (dirname, - ((flags & (GLOB_ERR | GLOB_NOCHECK | GLOB_NOESCAPE + ((flags & (GLOB_ERR | GLOB_NOESCAPE | GLOB_ALTDIRFUNC)) | GLOB_NOSORT | GLOB_ONLYDIR), errfunc, &dirs); if (status != 0) - return status; + { + if ((flags & GLOB_NOCHECK) == 0 || status != GLOB_NOMATCH) + return status; + goto no_matches; + } /* We have successfully globbed the preceding directory name. For each name we found, call glob_in_dir on it and FILENAME, @@ -811,6 +923,7 @@ glob (pattern, flags, errfunc, pglob) flag was set we must return the input pattern itself. */ if (pglob->gl_pathc + pglob->gl_offs == oldcount) { + no_matches: /* No matches. */ if (flags & GLOB_NOCHECK) { @@ -854,10 +967,44 @@ glob (pattern, flags, errfunc, pglob) else { int old_pathc = pglob->gl_pathc; + int orig_flags = flags; + if (meta & 2) + { + char *p = strchr (dirname, '\\'), *q; + /* We need to unescape the dirname string. It is certainly + allocated by alloca, as otherwise filename would be NULL + or dirname wouldn't contain backslashes. */ + q = p; + do + { + if (*p == '\\') + { + *q = *++p; + --dirlen; + } + else + *q = *p; + ++q; + } + while (*p++ != '\0'); + dirname_modified = 1; + } + if (dirname_modified) + flags &= ~(GLOB_NOCHECK | GLOB_NOMAGIC); status = glob_in_dir (filename, dirname, flags, errfunc, pglob); if (status != 0) - return status; + { + if (status == GLOB_NOMATCH && flags != orig_flags + && pglob->gl_pathc + pglob->gl_offs == oldcount) + { + /* Make sure globfree (&dirs); is a nop. */ + dirs.gl_pathv = NULL; + flags = orig_flags; + goto no_matches; + } + return status; + } if (dirlen > 0) { @@ -1015,15 +1162,13 @@ prefix_array (const char *dirname, char **array, size_t n) /* We must not compile this function twice. */ #if !defined _LIBC || !defined NO_GLOB_PATTERN_P -/* Return nonzero if PATTERN contains any metacharacters. - Metacharacters can be quoted with backslashes if QUOTE is nonzero. */ int -__glob_pattern_p (pattern, quote) +__glob_pattern_type (pattern, quote) const char *pattern; int quote; { register const char *p; - int open = 0; + int ret = 0; for (p = pattern; *p != '\0'; ++p) switch (*p) @@ -1033,21 +1178,35 @@ __glob_pattern_p (pattern, quote) return 1; case '\\': - if (quote && p[1] != '\0') - ++p; + if (quote) + { + if (p[1] != '\0') + ++p; + ret |= 2; + } break; case '[': - open = 1; + ret |= 4; break; case ']': - if (open) + if (ret & 4) return 1; break; } - return 0; + return ret; +} + +/* Return nonzero if PATTERN contains any metacharacters. + Metacharacters can be quoted with backslashes if QUOTE is nonzero. */ +int +__glob_pattern_p (pattern, quote) + const char *pattern; + int quote; +{ + return __glob_pattern_type (pattern, quote) == 1; } # ifdef _LIBC weak_alias (__glob_pattern_p, glob_pattern_p) @@ -1109,7 +1268,7 @@ glob_in_dir (const char *pattern, const char *directory, int flags, init_names.next = NULL; init_names.count = INITIAL_COUNT; - meta = __glob_pattern_p (pattern, !(flags & GLOB_NOESCAPE)); + meta = __glob_pattern_type (pattern, !(flags & GLOB_NOESCAPE)); if (meta == 0 && (flags & (GLOB_NOCHECK|GLOB_NOMAGIC))) { /* We need not do any tests. The PATTERN contains no meta @@ -1117,8 +1276,7 @@ glob_in_dir (const char *pattern, const char *directory, int flags, result will always contain exactly one name. */ flags |= GLOB_NOCHECK; } - else if (meta == 0 && - ((flags & GLOB_NOESCAPE) || strchr (pattern, '\\') == NULL)) + else if (meta == 0) { /* Since we use the normal file functions we can also use stat() to verify the file is there. */ @@ -1139,119 +1297,104 @@ glob_in_dir (const char *pattern, const char *directory, int flags, } else { - if (pattern[0] == '\0') + stream = ((flags & GLOB_ALTDIRFUNC) + ? (*pglob->gl_opendir) (directory) + : opendir (directory)); + if (stream == NULL) { - /* This is a special case for matching directories like in - "*a/". */ - names->name[cur] = (char *) malloc (1); - if (names->name[cur] == NULL) - goto memory_error; - *names->name[cur++] = '\0'; - nfound = 1; - meta = 0; + if (errno != ENOTDIR + && ((errfunc != NULL && (*errfunc) (directory, errno)) + || (flags & GLOB_ERR))) + return GLOB_ABORTED; } else { - stream = ((flags & GLOB_ALTDIRFUNC) - ? (*pglob->gl_opendir) (directory) - : opendir (directory)); - if (stream == NULL) - { - if (errno != ENOTDIR - && ((errfunc != NULL && (*errfunc) (directory, errno)) - || (flags & GLOB_ERR))) - return GLOB_ABORTED; - meta = 0; - } - else - { - int fnm_flags = ((!(flags & GLOB_PERIOD) ? FNM_PERIOD : 0) - | ((flags & GLOB_NOESCAPE) ? FNM_NOESCAPE : 0) + int fnm_flags = ((!(flags & GLOB_PERIOD) ? FNM_PERIOD : 0) + | ((flags & GLOB_NOESCAPE) ? FNM_NOESCAPE : 0) #if defined _AMIGA || defined VMS - | FNM_CASEFOLD + | FNM_CASEFOLD #endif - ); - flags |= GLOB_MAGCHAR; + ); + flags |= GLOB_MAGCHAR; - while (1) - { - const char *name; - size_t len; + while (1) + { + const char *name; + size_t len; #if defined _LIBC && !defined COMPILE_GLOB64 - struct dirent64 *d; - union - { - struct dirent64 d64; - char room [offsetof (struct dirent64, d_name[0]) - + NAME_MAX + 1]; - } - d64buf; + struct dirent64 *d; + union + { + struct dirent64 d64; + char room [offsetof (struct dirent64, d_name[0]) + + NAME_MAX + 1]; + } + d64buf; - if (flags & GLOB_ALTDIRFUNC) + if (flags & GLOB_ALTDIRFUNC) + { + struct dirent *d32 = (*pglob->gl_readdir) (stream); + if (d32 != NULL) { - struct dirent *d32 = (*pglob->gl_readdir) (stream); - if (d32 != NULL) - { - CONVERT_DIRENT_DIRENT64 (&d64buf.d64, d32); - d = &d64buf.d64; - } - else - d = NULL; + CONVERT_DIRENT_DIRENT64 (&d64buf.d64, d32); + d = &d64buf.d64; } else - d = __readdir64 (stream); + d = NULL; + } + else + d = __readdir64 (stream); #else - struct dirent *d = ((flags & GLOB_ALTDIRFUNC) - ? ((struct dirent *) - (*pglob->gl_readdir) (stream)) - : __readdir (stream)); + struct dirent *d = ((flags & GLOB_ALTDIRFUNC) + ? ((struct dirent *) + (*pglob->gl_readdir) (stream)) + : __readdir (stream)); #endif - if (d == NULL) - break; - if (! REAL_DIR_ENTRY (d)) - continue; + if (d == NULL) + break; + if (! REAL_DIR_ENTRY (d)) + continue; - /* If we shall match only directories use the information - provided by the dirent call if possible. */ - if ((flags & GLOB_ONLYDIR) && !DIRENT_MIGHT_BE_DIR (d)) - continue; + /* If we shall match only directories use the information + provided by the dirent call if possible. */ + if ((flags & GLOB_ONLYDIR) && !DIRENT_MIGHT_BE_DIR (d)) + continue; - name = d->d_name; + name = d->d_name; - if (fnmatch (pattern, name, fnm_flags) == 0) + if (fnmatch (pattern, name, fnm_flags) == 0) + { + /* If the file we found is a symlink we have to + make sure the target file exists. */ + if (!DIRENT_MIGHT_BE_SYMLINK (d) + || link_exists_p (directory, dirlen, name, pglob, + flags)) { - /* If the file we found is a symlink we have to - make sure the target file exists. */ - if (!DIRENT_MIGHT_BE_SYMLINK (d) - || link_exists_p (directory, dirlen, name, pglob, - flags)) + if (cur == names->count) { - if (cur == names->count) - { - struct globnames *newnames; - size_t count = names->count * 2; - size_t size = (sizeof (struct globnames) - + ((count - INITIAL_COUNT) - * sizeof (char *))); - allocasize += size; - if (__libc_use_alloca (allocasize)) - newnames = names_alloca = __alloca (size); - else if ((newnames = malloc (size)) - == NULL) - goto memory_error; - newnames->count = count; - newnames->next = names; - names = newnames; - cur = 0; - } - len = NAMLEN (d); - names->name[cur] = (char *) malloc (len + 1); - if (names->name[cur] == NULL) + struct globnames *newnames; + size_t count = names->count * 2; + size_t size = (sizeof (struct globnames) + + ((count - INITIAL_COUNT) + * sizeof (char *))); + allocasize += size; + if (__libc_use_alloca (allocasize)) + newnames = names_alloca = __alloca (size); + else if ((newnames = malloc (size)) + == NULL) goto memory_error; - *((char *) mempcpy (names->name[cur++], name, len)) - = '\0'; - ++nfound; + newnames->count = count; + newnames->next = names; + names = newnames; + cur = 0; } + len = NAMLEN (d); + names->name[cur] = (char *) malloc (len + 1); + if (names->name[cur] == NULL) + goto memory_error; + *((char *) mempcpy (names->name[cur++], name, len)) + = '\0'; + ++nfound; } } } diff --git a/posix/globtest.sh b/posix/globtest.sh index 09af6f7..ad47508 100755 --- a/posix/globtest.sh +++ b/posix/globtest.sh @@ -706,6 +706,77 @@ if test $failed -ne 0; then result=1 fi +mkdir $testdir/'dir3*' +echo 1 > $testdir/'dir3*'/file1 +mkdir $testdir/'dir4[a' +echo 2 > $testdir/'dir4[a'/file1 +echo 3 > $testdir/'dir4[a'/file2 +mkdir $testdir/'dir5[ab]' +echo 4 > $testdir/'dir5[ab]'/file1 +echo 5 > $testdir/'dir5[ab]'/file2 +mkdir $testdir/dir6 +echo 6 > $testdir/dir6/'file1[a' +echo 7 > $testdir/dir6/'file1[ab]' +failed=0 +v=`${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ + ${common_objpfx}posix/globtest "$testdir" 'dir3\*/file2'` +test "$v" != 'GLOB_NOMATCH' && echo "$v" >> $logfile && failed=1 +${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ +${common_objpfx}posix/globtest -c "$testdir" \ +'dir3\*/file1' 'dir3\*/file2' 'dir1/file\1_1' 'dir1/file\1_9' \ +'dir2\/' 'nondir\/' 'dir4[a/fil*1' 'di*r4[a/file2' 'dir5[ab]/file[12]' \ +'dir6/fil*[a' 'dir*6/file1[a' 'dir6/fi*l[ab]' 'dir*6/file1[ab]' | +sort > $testout +cat <<"EOF" | $CMP - $testout >> $logfile || failed=1 +`di*r4[a/file2' +`dir*6/file1[a' +`dir*6/file1[ab]' +`dir1/file1_1' +`dir1/file\1_9' +`dir2/' +`dir3*/file1' +`dir3\*/file2' +`dir4[a/fil*1' +`dir5[ab]/file[12]' +`dir6/fi*l[ab]' +`dir6/fil*[a' +`nondir\/' +EOF +HOME="$testdir" \ +${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ +${common_objpfx}posix/globtest -ct "$testdir" \ +'~/dir1/file1_1' '~/dir1/file1_9' '~/dir3\*/file1' '~/dir3\*/file2' \ +'~\/dir1/file1_2' | +sort > $testout +cat <<EOF | $CMP - $testout >> $logfile || failed=1 +\`$testdir/dir1/file1_1' +\`$testdir/dir1/file1_2' +\`$testdir/dir3*/file1' +\`~/dir1/file1_9' +\`~/dir3\\*/file2' +EOF +if eval test -d ~"$USER"/; then + user=`echo "$USER" | sed -n -e 's/^\([^\\]\)\([^\\][^\\]*\)$/~\1\\\\\2/p'` + if test -n "$user"; then + ${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ + ${common_objpfx}posix/globtest -ctq "$testdir" "$user/" | + sort > $testout + eval echo ~$USER/ | $CMP - $testout >> $logfile || failed=1 + ${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ + ${common_objpfx}posix/globtest -ctq "$testdir" "$user\\/" | + sort > $testout + eval echo ~$USER/ | $CMP - $testout >> $logfile || failed=1 + ${elf_objpfx}${rtld_installed_name} --library-path ${library_path} \ + ${common_objpfx}posix/globtest -ctq "$testdir" "$user" | + sort > $testout + eval echo ~$USER | $CMP - $testout >> $logfile || failed=1 + fi +fi +if test $failed -ne 0; then + echo "Escape tests failed" >> $logfile + result=1 +fi + if test $result -eq 0; then chmod 777 $testdir/noread rm -fr $testdir $testout |