diff --git i/dir.c w/dir.c index f9de4a3..1367ed0 100644 --- i/dir.c +++ w/dir.c @@ -897,11 +897,16 @@ check_dirname(volatile VALUE *dir) { VALUE d = *dir; char *path, *pend; + long len; + rb_encoding *enc; rb_secure(2); FilePathValue(d); - path = RSTRING_PTR(d); - if (path && *(pend = rb_path_end(rb_path_skip_prefix(path)))) { + enc = rb_enc_get(d); + RSTRING_GETMEM(d, path, len); + pend = path + len; + pend = rb_enc_path_end(rb_enc_path_skip_prefix(path, pend, enc), pend, enc); + if (pend - path < len) { d = rb_str_subseq(d, 0, pend - path); } *dir = rb_str_encode_ospath(d); @@ -1481,7 +1486,7 @@ ruby_glob0(const char *path, int flags, ruby_glob_func *func, VALUE arg, rb_enco start = root = path; flags |= FNM_SYSCASE; #if defined DOSISH - root = rb_path_skip_prefix(root); + root = rb_enc_path_skip_prefix(root, root + strlen(root), enc); #endif if (root && *root == '/') root++; diff --git i/ext/pathname/pathname.c w/ext/pathname/pathname.c index bde58a3..84a2b5d 100644 --- i/ext/pathname/pathname.c +++ w/ext/pathname/pathname.c @@ -1,4 +1,5 @@ #include "ruby.h" +#include "ruby/encoding.h" static VALUE rb_cPathname; static ID id_at_path, id_to_path; @@ -184,15 +185,15 @@ path_sub_ext(VALUE self, VALUE repl) StringValue(repl); p = RSTRING_PTR(str); - ext = ruby_find_extname(p, &extlen); + extlen = RSTRING_LEN(str); + ext = ruby_enc_find_extname(p, &extlen, rb_enc_get(str)); if (ext == NULL) { ext = p + RSTRING_LEN(str); } else if (extlen <= 1) { ext += extlen; } - str2 = rb_str_dup(str); - rb_str_resize(str2, ext-p); + str2 = rb_str_subseq(str, 0, ext-p); rb_str_append(str2, repl); OBJ_INFECT(str2, str); return rb_class_new_instance(1, &str2, rb_obj_class(self)); diff --git i/file.c w/file.c index 15d594a..04e78b7 100644 --- i/file.c +++ w/file.c @@ -2438,6 +2438,8 @@ rb_file_s_symlink(VALUE klass, VALUE from, VALUE to) #endif #ifdef HAVE_READLINK +static VALUE rb_readlink(VALUE path); + /* * call-seq: * File.readlink(link_name) -> file_name @@ -2452,6 +2454,12 @@ rb_file_s_symlink(VALUE klass, VALUE from, VALUE to) static VALUE rb_file_s_readlink(VALUE klass, VALUE path) { + return rb_readlink(path); +} + +static VALUE +rb_readlink(VALUE path) +{ char *buf; int size = 100; ssize_t rv; @@ -2617,9 +2625,8 @@ static const char file_alt_separator[] = {FILE_ALT_SEPARATOR, '\0'}; #define istrailinggarbage(x) 0 #endif -#ifndef CharNext /* defined as CharNext[AW] on Windows. */ -# define CharNext(p) ((p) + 1) -#endif +#define Next(p, e, enc) ((p) + rb_enc_mbclen((p), (e), (enc))) +#define Inc(p, e, enc) ((p) = Next((p), (e), (enc))) #if defined(DOSISH_UNC) #define has_unc(buf) (isdirsep((buf)[0]) && isdirsep((buf)[1])) @@ -2681,40 +2688,40 @@ not_same_drive(VALUE path, int drive) #endif static inline char * -skiproot(const char *path) +skiproot(const char *path, const char *end, rb_encoding *enc) { #ifdef DOSISH_DRIVE_LETTER - if (has_drive_letter(path)) path += 2; + if (path + 2 <= end && has_drive_letter(path)) path += 2; #endif - while (isdirsep(*path)) path++; + while (path < end && isdirsep(*path)) path++; return (char *)path; } -#define nextdirsep rb_path_next +#define nextdirsep rb_enc_path_next char * -rb_path_next(const char *s) +rb_enc_path_next(const char *s, const char *e, rb_encoding *enc) { - while (*s && !isdirsep(*s)) { - s = CharNext(s); + while (s < e && !isdirsep(*s)) { + Inc(s, e, enc); } return (char *)s; } #if defined(DOSISH_UNC) || defined(DOSISH_DRIVE_LETTER) -#define skipprefix rb_path_skip_prefix +#define skipprefix rb_enc_path_skip_prefix #else -#define skipprefix(path) (path) +#define skipprefix(path, end, enc) (path) #endif char * -rb_path_skip_prefix(const char *path) +rb_enc_path_skip_prefix(const char *path, const char *end, rb_encoding *enc) { #if defined(DOSISH_UNC) || defined(DOSISH_DRIVE_LETTER) #ifdef DOSISH_UNC - if (isdirsep(path[0]) && isdirsep(path[1])) { + if (path + 2 <= end && isdirsep(path[0]) && isdirsep(path[1])) { path += 2; - while (isdirsep(*path)) path++; - if (*(path = nextdirsep(path)) && path[1] && !isdirsep(path[1])) - path = nextdirsep(path + 1); + while (path < end && isdirsep(*path)) path++; + if ((path = rb_enc_path_next(path, end, enc)) < end && path[0] && path[1] && !isdirsep(path[1])) + path = rb_enc_path_next(path + 1, end, enc); return (char *)path; } #endif @@ -2727,78 +2734,78 @@ rb_path_skip_prefix(const char *path) } static inline char * -skipprefixroot(const char *path) +skipprefixroot(const char *path, const char *end, rb_encoding *enc) { #if defined(DOSISH_UNC) || defined(DOSISH_DRIVE_LETTER) - char *p = skipprefix(path); + char *p = skipprefix(path, end, enc); while (isdirsep(*p)) p++; return p; #else - return skiproot(path); + return skiproot(path, end, enc); #endif } -#define strrdirsep rb_path_last_separator +#define strrdirsep rb_enc_path_last_separator char * -rb_path_last_separator(const char *path) +rb_enc_path_last_separator(const char *path, const char *end, rb_encoding *enc) { char *last = NULL; - while (*path) { + while (path < end) { if (isdirsep(*path)) { const char *tmp = path++; - while (isdirsep(*path)) path++; - if (!*path) break; + while (path < end && isdirsep(*path)) path++; + if (path >= end) break; last = (char *)tmp; } else { - path = CharNext(path); + Inc(path, end, enc); } } return last; } static char * -chompdirsep(const char *path) +chompdirsep(const char *path, const char *end, rb_encoding *enc) { while (*path) { if (isdirsep(*path)) { const char *last = path++; - while (isdirsep(*path)) path++; - if (!*path) return (char *)last; + while (path < end && isdirsep(*path)) path++; + if (path >= end) return (char *)last; } else { - path = CharNext(path); + Inc(path, end, enc); } } return (char *)path; } char * -rb_path_end(const char *path) +rb_enc_path_end(const char *path, const char *end, rb_encoding *enc) { - if (isdirsep(*path)) path++; - return chompdirsep(path); + if (path < end && isdirsep(*path)) path++; + return chompdirsep(path, end, enc); } #if USE_NTFS static char * -ntfs_tail(const char *path) +ntfs_tail(const char *path, const char *end, rb_encoding *enc) { - while (*path == '.') path++; - while (*path && *path != ':') { + while (path < end && *path == '.') path++; + while (path < end && *path != ':') { if (istrailinggarbage(*path)) { const char *last = path++; - while (istrailinggarbage(*path)) path++; - if (!*path || *path == ':') return (char *)last; + while (path < end && istrailinggarbage(*path)) path++; + if (path >= end || *path == ':') return (char *)last; } else if (isdirsep(*path)) { const char *last = path++; - while (isdirsep(*path)) path++; - if (!*path) return (char *)last; + while (path < end && isdirsep(*path)) path++; + if (path >= end) return (char *)last; if (*path == ':') path++; } else { - path = CharNext(path); + Inc(path, end, enc); } } return (char *)path; @@ -2827,9 +2834,10 @@ rb_home_dir(const char *user, VALUE result) const char *dir; char *buf; #if defined DOSISH || defined __CYGWIN__ - char *p; + char *p, *bend; #endif long dirlen; + rb_encoding *enc; if (!user || !*user) { if (!(dir = getenv("HOME"))) { @@ -2848,32 +2856,36 @@ rb_home_dir(const char *user, VALUE result) } dirlen = strlen(pwPtr->pw_dir); rb_str_resize(result, dirlen); - strcpy(buf = RSTRING_PTR(result), pwPtr->pw_dir); + memcpy(buf = RSTRING_PTR(result), pwPtr->pw_dir, dirlen + 1); endpwent(); #else return Qnil; #endif } + enc = rb_filesystem_encoding(); + rb_enc_associate(result, enc); #if defined DOSISH || defined __CYGWIN__ - for (p = buf; *p; p = CharNext(p)) { + for (bend = (p = buf) + dirlen; p < bend; Inc(p, bend, enc)) { if (*p == '\\') { *p = '/'; } } #endif - rb_enc_associate_index(result, rb_filesystem_encindex()); return result; } static VALUE file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) { - const char *s, *b; + const char *s, *b, *fend; char *buf, *p, *pend, *root; size_t buflen, dirlen, bdiff; int tainted; + rb_encoding *enc, *fsenc = rb_filesystem_encoding(); s = StringValuePtr(fname); + fend = s + RSTRING_LEN(fname); + enc = rb_enc_get(fname); BUFINIT(); tainted = OBJ_TAINTED(fname); @@ -2887,7 +2899,7 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) if (*++s) ++s; } else { - s = nextdirsep(b = s); + s = nextdirsep(b = s, fend, enc); userlen = s - b; BUFCHECK(bdiff + userlen >= buflen); memcpy(p, b, userlen); @@ -2940,11 +2952,11 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) BUFCHECK(dirlen > buflen); strcpy(buf, dir); xfree(dir); - rb_enc_associate_index(result, rb_filesystem_encindex()); + rb_enc_associate(result, enc = fsenc); } else - rb_enc_associate(result, rb_enc_check(result, fname)); - p = chompdirsep(skiproot(buf)); + rb_enc_associate(result, enc = rb_enc_check(result, fname)); + p = chompdirsep(skiproot(buf, pend, enc), pend, enc); s += 2; } } @@ -2957,23 +2969,34 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) } else { char *dir = my_getcwd(); + char *cwdp; + VALUE dirname = Qnil; tainted = 1; dirlen = strlen(dir); + cwdp = dir; + if (enc != fsenc) { + rb_encoding *direnc = rb_enc_check(fname, dirname = rb_enc_str_new(dir, dirlen, fsenc)); + if (direnc != fsenc) { + dirname = rb_str_conv_enc(dirname, fsenc, direnc); + RSTRING_GETMEM(dirname, cwdp, dirlen); + } + } + rb_enc_associate(result, enc); BUFCHECK(dirlen > buflen); - strcpy(buf, dir); + strcpy(buf, cwdp); xfree(dir); - rb_enc_associate_index(result, rb_filesystem_encindex()); + if (!NIL_P(dirname)) rb_str_resize(dirname, 0); } #if defined DOSISH || defined __CYGWIN__ if (isdirsep(*s)) { /* specified full path, but not drive letter nor UNC */ /* we need to get the drive letter or UNC share name */ - p = skipprefix(buf); + p = skipprefix(buf, pend, enc); } else #endif - p = chompdirsep(skiproot(buf)); + p = chompdirsep(skiproot(buf, pend, enc), pend, enc); } else { size_t len; @@ -2997,7 +3020,7 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) rb_str_set_len(result, p-buf+1); BUFCHECK(bdiff + 1 >= buflen); p[1] = 0; - root = skipprefix(buf); + root = skipprefix(buf, p+1, enc); b = s; while (*s) { @@ -3013,7 +3036,7 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) /* We must go back to the parent */ char *n; *p = '\0'; - if (!(n = strrdirsep(root))) { + if (!(n = strrdirsep(root, p, enc))) { *p = '/'; } else { @@ -3043,7 +3066,7 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) --s; case ' ': { const char *e = s; - while (istrailinggarbage(*s)) s++; + while (s < fend && istrailinggarbage(*s)) s++; if (!*s) { s = e; goto endpath; @@ -3068,7 +3091,7 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) b = ++s; break; default: - s = CharNext(s); + Inc(s, fend, enc); break; } } @@ -3093,14 +3116,18 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) BUFCHECK(bdiff + (s-b) >= buflen); memcpy(++p, b, s-b); p += s-b; + rb_str_set_len(result, p-buf); } - if (p == skiproot(buf) - 1) p++; + if (p == skiproot(buf, p + !!*p, enc) - 1) p++; #if USE_NTFS *p = '\0'; - if ((s = strrdirsep(b = buf)) != 0 && !strpbrk(s, "*?")) { + if ((s = strrdirsep(b = buf, p, enc)) != 0 && !strpbrk(s, "*?")) { + VALUE tmp, v; size_t len; - WIN32_FIND_DATA wfd; + rb_encoding *enc; + WCHAR *wstr; + WIN32_FIND_DATAW wfd; HANDLE h; #ifdef __CYGWIN__ #ifdef HAVE_CYGWIN_CONV_PATH @@ -3150,21 +3177,43 @@ file_expand_path(VALUE fname, VALUE dname, int abs_mode, VALUE result) } *p = '/'; #endif - h = FindFirstFile(b, &wfd); + rb_str_set_len(result, p - buf + strlen(p)); + enc = rb_enc_get(result); + tmp = result; + if (enc != rb_utf8_encoding() && rb_enc_str_coderange(result) != ENC_CODERANGE_7BIT) { + tmp = rb_str_encode_ospath(result); + } + len = MultiByteToWideChar(CP_UTF8, 0, RSTRING_PTR(tmp), -1, NULL, 0); + wstr = ALLOCV_N(WCHAR, v, len); + MultiByteToWideChar(CP_UTF8, 0, RSTRING_PTR(tmp), -1, wstr, len); + if (tmp != result) rb_str_resize(tmp, 0); + h = FindFirstFileW(wstr, &wfd); + ALLOCV_END(v); if (h != INVALID_HANDLE_VALUE) { + size_t wlen; FindClose(h); - len = strlen(wfd.cFileName); + len = lstrlenW(wfd.cFileName); #ifdef __CYGWIN__ if (lnk_added && len > 4 && - STRCASECMP(wfd.cFileName + len - 4, ".lnk") == 0) { - wfd.cFileName[len -= 4] = '\0'; + wcsicmp(wfd.cFileName + len - 4, L".lnk") == 0) { + wfd.cFileName[len -= 4] = L'\0'; } #else p = (char *)s; #endif ++p; + wlen = (int)len; + len = WideCharToMultiByte(CP_UTF8, 0, wfd.cFileName, wlen, NULL, 0, NULL, NULL); BUFCHECK(bdiff + len >= buflen); - memcpy(p, wfd.cFileName, len + 1); + WideCharToMultiByte(CP_UTF8, 0, wfd.cFileName, wlen, p, len + 1, NULL, NULL); + if (tmp != result) { + rb_str_buf_cat(tmp, p, len); + tmp = rb_str_encode(tmp, rb_enc_from_encoding(enc), 0, Qnil); + len = RSTRING_LEN(tmp); + BUFCHECK(bdiff + len >= buflen); + memcpy(p, RSTRING_PTR(tmp), len); + rb_str_resize(tmp, 0); + } p += len; } #ifdef __CYGWIN__ @@ -3265,23 +3314,27 @@ rb_file_s_absolute_path(int argc, VALUE *argv) } static void -realpath_rec(long *prefixlenp, VALUE *resolvedp, char *unresolved, VALUE loopcheck, int strict, int last) +realpath_rec(long *prefixlenp, VALUE *resolvedp, const char *unresolved, VALUE loopcheck, int strict, int last) { + const char *pend = unresolved + strlen(unresolved); + rb_encoding *enc = rb_enc_get(*resolvedp); ID resolving; CONST_ID(resolving, "resolving"); - while (*unresolved) { - char *testname = unresolved; - char *unresolved_firstsep = rb_path_next(unresolved); - long testnamelen = unresolved_firstsep - unresolved; - char *unresolved_nextname = unresolved_firstsep; - while (isdirsep(*unresolved_nextname)) unresolved_nextname++; + while (unresolved < pend) { + const char *testname = unresolved; + const char *unresolved_firstsep = rb_enc_path_next(unresolved, pend, enc); + long testnamelen = unresolved_firstsep - unresolved; + const char *unresolved_nextname = unresolved_firstsep; + while (unresolved_nextname < pend && isdirsep(*unresolved_nextname)) + unresolved_nextname++; unresolved = unresolved_nextname; if (testnamelen == 1 && testname[0] == '.') { } else if (testnamelen == 2 && testname[0] == '.' && testname[1] == '.') { if (*prefixlenp < RSTRING_LEN(*resolvedp)) { - char *resolved_names = RSTRING_PTR(*resolvedp) + *prefixlenp; - char *lastsep = rb_path_last_separator(resolved_names); + const char *resolved_str = RSTRING_PTR(*resolvedp); + const char *resolved_names = resolved_str + *prefixlenp; + const char *lastsep = strrdirsep(resolved_names, resolved_str + RSTRING_LEN(*resolvedp), enc); long len = lastsep ? lastsep - resolved_names : 0; rb_str_resize(*resolvedp, *prefixlenp + len); } @@ -3320,23 +3373,24 @@ realpath_rec(long *prefixlenp, VALUE *resolvedp, char *unresolved, VALUE loopche } #ifdef HAVE_READLINK if (S_ISLNK(sbuf.st_mode)) { - volatile VALUE link; - char *link_prefix, *link_names; + VALUE link; + const char *link_prefix, *link_names; long link_prefixlen; rb_hash_aset(loopcheck, testpath, ID2SYM(resolving)); - link = rb_file_s_readlink(rb_cFile, testpath); + link = rb_readlink(testpath); link_prefix = RSTRING_PTR(link); - link_names = skipprefixroot(link_prefix); - link_prefixlen = link_names - link_prefix; - if (link_prefixlen == 0) { - realpath_rec(prefixlenp, resolvedp, link_names, loopcheck, strict, *unresolved_firstsep == '\0'); - } - else { - *resolvedp = rb_str_new(link_prefix, link_prefixlen); - *prefixlenp = link_prefixlen; - realpath_rec(prefixlenp, resolvedp, link_names, loopcheck, strict, *unresolved_firstsep == '\0'); - } - rb_hash_aset(loopcheck, testpath, rb_str_dup_frozen(*resolvedp)); + link_names = skipprefixroot(link_prefix, link_prefix + RSTRING_LEN(link), rb_enc_get(link)); + link_prefixlen = link_names - link_prefix; + if (link_prefixlen > 0) { + rb_encoding *enc, *linkenc = rb_enc_get(link); + link = rb_str_subseq(link, 0, link_prefixlen); + enc = rb_enc_check(*resolvedp, link); + if (enc != linkenc) link = rb_str_conv_enc(link, linkenc, enc); + *resolvedp = link; + *prefixlenp = link_prefixlen; + } + realpath_rec(prefixlenp, resolvedp, link_names, loopcheck, strict, *unresolved_firstsep == '\0'); + rb_hash_aset(loopcheck, testpath, rb_str_dup_frozen(*resolvedp)); } else #endif @@ -3359,8 +3413,10 @@ rb_realpath_internal(VALUE basedir, VALUE path, int strict) VALUE loopcheck; volatile VALUE curdir = Qnil; + rb_encoding *enc; char *path_names = NULL, *basedir_names = NULL, *curdir_names = NULL; - char *ptr, *prefixptr = NULL; + char *ptr, *prefixptr = NULL, *pend; + long len; rb_secure(2); @@ -3372,34 +3428,33 @@ rb_realpath_internal(VALUE basedir, VALUE path, int strict) basedir = rb_str_dup_frozen(basedir); } - ptr = RSTRING_PTR(unresolved_path); - path_names = skipprefixroot(ptr); + RSTRING_GETMEM(unresolved_path, ptr, len); + path_names = skipprefixroot(ptr, ptr + len, rb_enc_get(unresolved_path)); if (ptr != path_names) { - resolved = rb_enc_str_new(ptr, path_names - ptr, - rb_enc_get(unresolved_path)); + resolved = rb_str_subseq(unresolved_path, 0, path_names - ptr); goto root_found; } if (!NIL_P(basedir)) { - ptr = RSTRING_PTR(basedir); - basedir_names = skipprefixroot(ptr); + RSTRING_GETMEM(basedir, ptr, len); + basedir_names = skipprefixroot(ptr, ptr + len, rb_enc_get(basedir)); if (ptr != basedir_names) { - resolved = rb_enc_str_new(ptr, basedir_names - ptr, - rb_enc_get(basedir)); - goto root_found; + resolved = rb_str_subseq(basedir, 0, basedir_names - ptr); + goto root_found; } } curdir = rb_dir_getwd(); - ptr = RSTRING_PTR(curdir); - curdir_names = skipprefixroot(ptr); - resolved = rb_enc_str_new(ptr, curdir_names - ptr, rb_enc_get(curdir)); + RSTRING_GETMEM(curdir, ptr, len); + curdir_names = skipprefixroot(ptr, ptr + len, rb_enc_get(curdir)); + resolved = rb_str_subseq(curdir, 0, curdir_names - ptr); root_found: - prefixptr = RSTRING_PTR(resolved); - prefixlen = RSTRING_LEN(resolved); - ptr = chompdirsep(prefixptr); - if (*ptr) { + RSTRING_GETMEM(resolved, prefixptr, prefixlen); + pend = prefixptr + prefixlen; + enc = rb_enc_get(resolved); + ptr = chompdirsep(prefixptr, pend, enc); + if (ptr < pend) { prefixlen = ++ptr - prefixptr; rb_str_set_len(resolved, prefixlen); } @@ -3408,7 +3463,7 @@ rb_realpath_internal(VALUE basedir, VALUE path, int strict) if (*prefixptr == FILE_ALT_SEPARATOR) { *prefixptr = '/'; } - prefixptr = CharNext(prefixptr); + Inc(prefixptr, pend, enc); } #endif @@ -3494,15 +3549,16 @@ rmext(const char *p, long l0, long l1, const char *e) } const char * -ruby_find_basename(const char *name, long *baselen, long *alllen) +ruby_enc_find_basename(const char *name, long *baselen, long *alllen, rb_encoding *enc) { - const char *p, *q, *e; + const char *p, *q, *e, *end; #if defined DOSISH_DRIVE_LETTER || defined DOSISH_UNC const char *root; #endif long f = 0, n = -1; - name = skipprefix(name); + end = name + *alllen; + name = skipprefix(name, end, enc); #if defined DOSISH_DRIVE_LETTER || defined DOSISH_UNC root = name; #endif @@ -3529,19 +3585,19 @@ ruby_find_basename(const char *name, long *baselen, long *alllen) #endif } else { - if (!(p = strrdirsep(name))) { + if (!(p = strrdirsep(name, end, enc))) { p = name; } else { while (isdirsep(*p)) p++; /* skip last / */ } #if USE_NTFS - n = ntfs_tail(p) - p; + n = ntfs_tail(p, end, enc) - p; #else - n = chompdirsep(p) - p; + n = chompdirsep(p, end, enc) - p; #endif for (q = p; q - p < n && *q == '.'; q++); - for (e = 0; q - p < n; q = CharNext(q)) { + for (e = 0; q - p < n; Inc(q, end, enc)) { if (*q == '.') e = q; } if (e) f = e - p; @@ -3575,6 +3631,7 @@ rb_file_s_basename(int argc, VALUE *argv) VALUE fname, fext, basename; const char *name, *p; long f, n; + rb_encoding *enc; if (rb_scan_args(argc, argv, "11", &fname, &fext) == 2) { rb_encoding *enc; @@ -3585,11 +3642,12 @@ rb_file_s_basename(int argc, VALUE *argv) } } FilePathStringValue(fname); - if (!NIL_P(fext)) rb_enc_check(fname, fext); - if (RSTRING_LEN(fname) == 0 || !*(name = RSTRING_PTR(fname))) + if (!NIL_P(fext)) enc = rb_enc_check(fname, fext); + else enc = rb_enc_get(fname); + if ((n = RSTRING_LEN(fname)) == 0 || !*(name = RSTRING_PTR(fname))) return rb_str_new_shared(fname); - p = ruby_find_basename(name, &f, &n); + p = ruby_enc_find_basename(name, &f, &n, enc); if (n >= 0) { if (NIL_P(fext) || !(f = rmext(p, f, n, StringValueCStr(fext)))) { f = n; @@ -3624,20 +3682,23 @@ rb_file_s_dirname(VALUE klass, VALUE fname) VALUE rb_file_dirname(VALUE fname) { - const char *name, *root, *p; + const char *name, *root, *p, *end; VALUE dirname; + rb_encoding *enc; FilePathStringValue(fname); name = StringValueCStr(fname); - root = skiproot(name); + end = name + RSTRING_LEN(fname); + enc = rb_enc_get(fname); + root = skiproot(name, end, enc); #ifdef DOSISH_UNC if (root > name + 1 && isdirsep(*name)) - root = skipprefix(name = root - 2); + root = skipprefix(name = root - 2, end, enc); #else if (root > name + 1) name = root - 1; #endif - p = strrdirsep(root); + p = strrdirsep(root, end, enc); if (!p) { p = root; } @@ -3645,7 +3706,7 @@ rb_file_dirname(VALUE fname) return rb_usascii_str_new2("."); #ifdef DOSISH_DRIVE_LETTER if (has_drive_letter(name) && isdirsep(*(name + 2))) { - const char *top = skiproot(name + 2); + const char *top = skiproot(name + 2, end, enc); dirname = rb_str_new(name, 3); rb_str_cat(dirname, top, p - top); } @@ -3674,11 +3735,11 @@ rb_file_dirname(VALUE fname) * */ const char * -ruby_find_extname(const char *name, long *len) +ruby_enc_find_extname(const char *name, long *len, rb_encoding *enc) { - const char *p, *e; + const char *p, *e, *end = name + (len ? *len : (long)strlen(name)); - p = strrdirsep(name); /* get the last path component */ + p = strrdirsep(name, end, enc); /* get the last path component */ if (!p) p = name; else @@ -3711,7 +3772,7 @@ ruby_find_extname(const char *name, long *len) #endif else if (isdirsep(*p)) break; - p = CharNext(p); + Inc(p, end, enc); } if (len) { @@ -3749,11 +3810,11 @@ rb_file_s_extname(VALUE klass, VALUE fname) FilePathStringValue(fname); name = StringValueCStr(fname); - e = ruby_find_extname(name, &len); + len = RSTRING_LEN(fname); + e = ruby_enc_find_extname(name, &len, rb_enc_get(fname)); if (len <= 1) return rb_str_new(0, 0); - extname = rb_str_new(e, len); /* keep the dot, too! */ - rb_enc_copy(extname, fname); + extname = rb_str_subseq(fname, e - name, len); /* keep the dot, too! */ OBJ_INFECT(extname, fname); return extname; } @@ -3816,8 +3877,9 @@ rb_file_join(VALUE ary, VALUE sep) len = 1; for (i=0; iklass = 0; + len = 0; for (i=0; iklass) rb_raise(rb_eRuntimeError, "File.join reentered"); } - name = StringValueCStr(result); if (i == 0) { rb_enc_copy(result, tmp); } else if (!NIL_P(sep)) { - tail = chompdirsep(name); + name = RSTRING_PTR(result); + tail = chompdirsep(name + len, name + RSTRING_LEN(result), rb_enc_get(result)); if (RSTRING_PTR(tmp) && isdirsep(RSTRING_PTR(tmp)[0])) { rb_str_set_len(result, tail - name); } @@ -3864,6 +3929,7 @@ rb_file_join(VALUE ary, VALUE sep) } rb_str_buf_append(result, tmp); } + RBASIC(result)->klass = rb_cString; return result; } @@ -4990,6 +5056,8 @@ path_check_0(VALUE path, int execpath) { struct stat st; const char *p0 = StringValueCStr(path); + const char *e0; + rb_encoding *enc; char *p = 0, *s; if (!rb_is_absolute_path(p0)) { @@ -5004,6 +5072,8 @@ path_check_0(VALUE path, int execpath) path = newpath; p0 = RSTRING_PTR(path); } + e0 = p0 + RSTRING_LEN(path); + enc = rb_enc_get(path); for (;;) { #ifndef S_IWOTH # define S_IWOTH 002 @@ -5020,10 +5090,11 @@ path_check_0(VALUE path, int execpath) RB_GC_GUARD(path); return 0; } - s = strrdirsep(p0); + s = strrdirsep(p0, e0, enc); if (p) *p = '/'; if (!s || s == p0) return 1; p = s; + e0 = p; *p = '\0'; } } diff --git i/include/ruby/encoding.h w/include/ruby/encoding.h index 058462f..f2dcebb 100644 --- i/include/ruby/encoding.h +++ w/include/ruby/encoding.h @@ -159,10 +159,12 @@ int rb_enc_codelen(int code, rb_encoding *enc); /* start, ptr, end, encoding -> prev_char */ #define rb_enc_prev_char(s,p,e,enc) ((char *)onigenc_get_prev_char_head((enc),(UChar*)(s),(UChar*)(p),(UChar*)(e))) -/* start, ptr, end, encoding -> next_char */ +/* ptr, end, encoding -> next_char */ +#define rb_enc_next_char(p,e,enc) ((char *)onigenc_get_prev_char_head((enc),(UChar*)(p),(UChar*)(e))) #define rb_enc_left_char_head(s,p,e,enc) ((char *)onigenc_get_left_adjust_char_head((enc),(UChar*)(s),(UChar*)(p),(UChar*)(e))) #define rb_enc_right_char_head(s,p,e,enc) ((char *)onigenc_get_right_adjust_char_head((enc),(UChar*)(s),(UChar*)(p),(UChar*)(e))) #define rb_enc_step_back(s,p,e,n,enc) ((char *)onigenc_step_back((enc),(UChar*)(s),(UChar*)(p),(UChar*)(e),(int)(n))) +#define rb_enc_step(p,e,n,enc) ((char *)onigenc_step((enc),(UChar*)(p),(UChar*)(e),(int)(n))) /* ptr, ptr, encoding -> newline_or_not */ #define rb_enc_is_newline(p,end,enc) ONIGENC_IS_MBC_NEWLINE((enc),(UChar*)(p),(UChar*)(end)) @@ -211,6 +213,12 @@ void rb_enc_set_default_external(VALUE encoding); void rb_enc_set_default_internal(VALUE encoding); VALUE rb_locale_charmap(VALUE klass); long rb_memsearch(const void*,long,const void*,long,rb_encoding*); +char *rb_enc_path_next(const char *,const char *,rb_encoding*); +char *rb_enc_path_skip_prefix(const char *,const char *,rb_encoding*); +char *rb_enc_path_last_separator(const char *,const char *,rb_encoding*); +char *rb_enc_path_end(const char *,const char *,rb_encoding*); +const char *ruby_enc_find_basename(const char *name, long *baselen, long *alllen, rb_encoding *enc); +const char *ruby_enc_find_extname(const char *name, long *len, rb_encoding *enc); RUBY_EXTERN VALUE rb_cEncoding; #define ENC_DUMMY_FLAG (1<<24) diff --git i/include/ruby/intern.h w/include/ruby/intern.h index fbc9c49..0546c15 100644 --- i/include/ruby/intern.h +++ w/include/ruby/intern.h @@ -426,8 +426,6 @@ char *rb_path_end(const char *); VALUE rb_file_directory_p(VALUE,VALUE); VALUE rb_str_encode_ospath(VALUE); int rb_is_absolute_path(const char *); -const char *ruby_find_basename(const char *name, long *baselen, long *alllen); -const char *ruby_find_extname(const char *name, long *len); /* gc.c */ void ruby_set_stack_size(size_t); NORETURN(void rb_memerror(void)); diff --git i/test/pathname/test_pathname.rb w/test/pathname/test_pathname.rb index 401bad2..dcbcd49 100644 --- i/test/pathname/test_pathname.rb +++ w/test/pathname/test_pathname.rb @@ -1,5 +1,3 @@ -#!/usr/bin/env ruby - require 'test/unit' require 'pathname' @@ -185,10 +183,8 @@ class TestPathname < Test::Unit::TestCase if DOSISH defassert(:del_trailing_separator, "a", "a\\") - require 'Win32API' - if Win32API.new('kernel32', 'GetACP', nil, 'L').call == 932 - defassert(:del_trailing_separator, "\225\\", "\225\\\\") # SJIS - end + defassert(:del_trailing_separator, "\225\\".force_encoding("cp932"), "\225\\\\".force_encoding("cp932")) + defassert(:del_trailing_separator, "\225".force_encoding("cp437"), "\225\\\\".force_encoding("cp437")) end def test_plus diff --git i/test/ruby/test_file_exhaustive.rb w/test/ruby/test_file_exhaustive.rb index c10b05a..a27c6ad 100644 --- i/test/ruby/test_file_exhaustive.rb +++ w/test/ruby/test_file_exhaustive.rb @@ -3,6 +3,8 @@ require "fileutils" require "tmpdir" class TestFileExhaustive < Test::Unit::TestCase + DRIVE = Dir.pwd[%r'\A(?:[a-z]:|//[^/]+/[^/]+)'i] + def assert_incompatible_encoding d = "\u{3042}\u{3044}".encode("utf-16le") assert_raise(Encoding::CompatibilityError) {yield d} @@ -400,13 +402,27 @@ class TestFileExhaustive < Test::Unit::TestCase assert_match(/\Ac:\//i, File.expand_path('c:foo', 'd:/bar')) assert_match(%r'\Ac:/bar/foo\z'i, File.expand_path('c:foo', 'c:/bar')) end - if drive = Dir.pwd[%r'\A(?:[a-z]:|//[^/]+/[^/]+)'i] + if DRIVE assert_match(%r"\Az:/foo\z"i, File.expand_path('/foo', "z:/bar")) assert_match(%r"\A//host/share/foo\z"i, File.expand_path('/foo', "//host/share/bar")) - assert_match(%r"\A#{drive}/foo\z"i, File.expand_path('/foo')) + assert_match(%r"\A#{DRIVE}/foo\z"i, File.expand_path('/foo')) else assert_equal("/foo", File.expand_path('/foo')) end + if Encoding.find("filesystem") == Encoding::CP1251 + a = "/\u3042\u3044\u3046\u3048\u304a".encode("cp932") + else + a = "/\u043f\u0440\u0438\u0432\u0435\u0442".encode("cp1251") + end + a = "C:" + a if DRIVE + assert_equal(a, File.expand_path(a)) + if DRIVE + a = "C:/\225\\\\" + [%W"cp437 C:/\225", %W"cp932 C:/\225\\"].each do |cp, expected| + assert_equal(expected.force_encoding(cp), File.expand_path(a.dup.force_encoding(cp)), cp) + end + end + assert_kind_of(String, File.expand_path("~")) if ENV["HOME"] assert_raise(ArgumentError) { File.expand_path("~foo_bar_baz_unknown_user_wahaha") } assert_raise(ArgumentError) { File.expand_path("~foo_bar_baz_unknown_user_wahaha", "/") } @@ -447,6 +463,12 @@ class TestFileExhaustive < Test::Unit::TestCase assert_equal(basename, File.basename(@file + ".", ".*")) assert_equal(basename, File.basename(@file + "::$DATA", ".*")) end + if DRIVE + a = "C:/\225\\\\" + [%W"cp437 \225", %W"cp932 \225\\"].each do |cp, expected| + assert_equal(expected.force_encoding(cp), File.basename(a.dup.force_encoding(cp)), cp) + end + end assert_incompatible_encoding {|d| File.basename(d)} assert_incompatible_encoding {|d| File.basename(d, ".*")}