inc/SafeFN.class.php

   1 <?php
   2
   3 /**
   4  * Class to safely store UTF-8 in a Filename
   5  *
   6  * Encodes a utf8 string using only the following characters 0-9a-z_.-%
   7  * characters 0-9a-z in the original string are preserved, "plain".
   8  * all other characters are represented in a substring that starts
   9  * with '%' are "converted".
  10  * The transition from converted substrings to plain characters is
  11  * marked with a '.'
  12  *
  13  * @author   Christopher Smith <chris@jalakai.co.uk>
  14  * @date     2010-04-02
  15  */
  16 class SafeFN {
  17
  18     // 'safe' characters are a superset of $plain, $pre_indicator and $post_indicator
  19     private static $plain = '-./[_0123456789abcdefghijklmnopqrstuvwxyz'; // these characters aren't converted
  20     private static $pre_indicator = '%';
  21     private static $post_indicator = ']';
  22
  23     /**
  24      * Convert an UTF-8 string to a safe ASCII String
  25      *
  26      *  conversion process
  27      *    - if codepoint is a plain or post_indicator character,
  28      *      - if previous character was "converted", append post_indicator to output, clear "converted" flag
  29      *      - append ascii byte for character to output
  30      *      (continue to next character)
  31      *
  32      *    - if codepoint is a pre_indicator character,
  33      *      - append ascii byte for character to output, set "converted" flag
  34      *      (continue to next character)
  35      *
  36      *    (all remaining characters)
  37      *    - reduce codepoint value for non-printable ASCII characters (0x00 - 0x1f).  Space becomes our zero.
  38      *    - convert reduced value to base36 (0-9a-z)
  39      *    - append $pre_indicator characater followed by base36 string to output, set converted flag
  40      *    (continue to next character)
  41      *
  42      * @param    string    $filename     a utf8 string, should only include printable characters - not 0x00-0x1f
  43      * @return   string    an encoded representation of $filename using only 'safe' ASCII characters
  44      *
  45      * @author   Christopher Smith <chris@jalakai.co.uk>
  46      */
  47     public static function encode($filename) {
  48         return self::unicodeToSafe(\dokuwiki\Utf8\Unicode::fromUtf8($filename));
  49     }
  50
  51     /**
  52      *  decoding process
  53      *    - split the string into substrings at any occurrence of pre or post indicator characters
  54      *    - check the first character of the substring
  55      *      - if its not a pre_indicator character
  56      *        - if previous character was converted, skip over post_indicator character
  57      *        - copy codepoint values of remaining characters to the output array
  58      *        - clear any converted flag
  59      *      (continue to next substring)
  60      *
  61      *     _ else (its a pre_indicator character)
  62      *       - if string length is 1, copy the post_indicator character to the output array
  63      *       (continue to next substring)
  64      *
  65      *       - else (string length > 1)
  66      *         - skip the pre-indicator character and convert remaining string from base36 to base10
  67      *         - increase codepoint value for non-printable ASCII characters (add 0x20)
  68      *         - append codepoint to output array
  69      *       (continue to next substring)
  70      *
  71      * @param    string    $filename     a 'safe' encoded ASCII string,
  72      * @return   string    decoded utf8 representation of $filename
  73      *
  74      * @author   Christopher Smith <chris@jalakai.co.uk>
  75      */
  76     public static function decode($filename) {
  77         return \dokuwiki\Utf8\Unicode::toUtf8(self::safeToUnicode(strtolower($filename)));
  78     }
  79
  80     public static function validatePrintableUtf8($printable_utf8) {
  81         return !preg_match('#[\x01-\x1f]#',$printable_utf8);
  82     }
  83
  84     public static function validateSafe($safe) {
  85         return !preg_match('#[^'.self::$plain.self::$post_indicator.self::$pre_indicator.']#',$safe);
  86     }
  87
  88     /**
  89      * convert an array of unicode codepoints into 'safe_filename' format
  90      *
  91      * @param    array  int    $unicode    an array of unicode codepoints
  92      * @return   string        the unicode represented in 'safe_filename' format
  93      *
  94      * @author   Christopher Smith <chris@jalakai.co.uk>
  95      */
  96     private static function unicodeToSafe($unicode) {
  97
  98         $safe = '';
  99         $converted = false;
 100
 101         foreach ($unicode as $codepoint) {
 102             if ($codepoint < 127 && (strpos(self::$plain.self::$post_indicator,chr($codepoint))!==false)) {
 103                 if ($converted) {
 104                     $safe .= self::$post_indicator;
 105                     $converted = false;
 106                 }
 107                 $safe .= chr($codepoint);
 108
 109             } else if ($codepoint == ord(self::$pre_indicator)) {
 110                 $safe .= self::$pre_indicator;
 111                 $converted = true;
 112             } else {
 113                 $safe .= self::$pre_indicator.base_convert((string)($codepoint-32),10,36);
 114                 $converted = true;
 115             }
 116         }
 117         if($converted) $safe .= self::$post_indicator;
 118         return $safe;
 119     }
 120
 121     /**
 122      * convert a 'safe_filename' string into an array of unicode codepoints
 123      *
 124      * @param   string         $safe     a filename in 'safe_filename' format
 125      * @return  array   int    an array of unicode codepoints
 126      *
 127      * @author   Christopher Smith <chris@jalakai.co.uk>
 128      */
 129     private static function safeToUnicode($safe) {
 130
 131         $unicode = array();
 132         $split = preg_split('#(?=['.self::$post_indicator.self::$pre_indicator.'])#',$safe,-1,PREG_SPLIT_NO_EMPTY);
 133
 134         $converted = false;
 135         foreach ($split as $sub) {
 136             $len = strlen($sub);
 137             if ($sub[0] != self::$pre_indicator) {
 138                 // plain (unconverted) characters, optionally starting with a post_indicator
 139                 // set initial value to skip any post_indicator
 140                 for ($i=($converted?1:0); $i < $len; $i++) {
 141                     $unicode[] = ord($sub[$i]);
 142                 }
 143                 $converted = false;
 144             } else if ($len==1) {
 145                 // a pre_indicator character in the real data
 146                 $unicode[] = ord($sub);
 147                 $converted = true;
 148             } else {
 149                 // a single codepoint in base36, adjusted for initial 32 non-printable chars
 150                 $unicode[] = 32 + (int)base_convert(substr($sub,1),36,10);
 151                 $converted = true;
 152             }
 153         }
 154
 155         return $unicode;
 156     }
 157
 158 }