#46990 [Opn]: Passing UTF8 strings to filesystem functions produce wrong filenames

[alex dot bazan at concatel dot com]

 ID:               46990
 User updated by:  alex dot bazan at concatel dot com
 Reported By:      alex dot bazan at concatel dot com
 Status:           Open
 Bug Type:         *Directory/Filesystem functions
 Operating System: win32 only - Windows XP
 PHP Version:      5.2.8
 New Comment:

For anyone stumbling upon this bug, i wrote a workaround building a
wrapper for filesystem functions. This workaround will will not fix the
issue completely but will work with latin compatible characters when
using UTF8 as the string encoding. People using other characters
(japanese, chinese, hebrew ...) will still have problems.


CODE:


/**
 * Wrapper for PHP filesystem functions
 */
class Fsw {

    /**
     * Returns the filename for use with filesystem functions
     */
    static function setName ($file) {
        if (DIRECTORY_SEPARATOR=="\\") {
            $file=utf8_decode($file);
        }
        return $file;
    }

    /**
     * Encondes the filename returned by filesystem functions
     */
    static function getName ($file) {
        if (DIRECTORY_SEPARATOR=="\\") {
            $file=utf8_encode($file);
        }
        return $file;
    }


    static function file_exists ($filename) {
        return file_exists(self::setName($filename));
    }

    static function fopen ($filename,$mode) {
        return fopen(self::setName($filename),$mode);
    }

    static function opendir ($dirname) {
        return opendir(self::setName($dirname));
    }

    static function readdir ($dir_handle) {
        $file=readdir($dir_handle);
        if ($file) {
            // check if file is found before converting it's
            // name or we will convert bool(false) to string
            $file=self::getName($file);
        }
        return $file;
    }

// just create any other filesystem function you might use in your code
here

}


Previous Comments:
------------------------------------------------------------------------

[2009-01-03 13:25:07] j...@php.net




------------------------------------------------------------------------

[2009-01-02 08:06:39] alex dot bazan at concatel dot com

The UTF string did not get saved correctly in the bug report. It was a
japanese string.

------------------------------------------------------------------------

[2009-01-02 08:03:37] alex dot bazan at concatel dot com

Description:
------------
Under Windows, when I use fpoen() or mkdir() with a UTF8 encoded
string, the file or directory name is not converted to the filesystem
encoding and thus the file or directory is created with a wrong file
name.

Reproduce code:
---------------
<?php mkdir('&#26085;&#26412;&#35486;'); ?>

Expected result:
----------------
Directory &#26085;&#26412;&#35486; should be created

Actual result:
--------------
Directory 日本語 is created


------------------------------------------------------------------------


-- 
Edit this bug report at http://bugs.php.net/?id=46990&edit=1