banana/utf8.php

   1 <?php
   2 /********************************************************************************
   3 * banana/utf8.php : utf8 to html entities
   4 * ---------------
   5 *
   6 * This file is part of the banana distribution
   7 * Copyright: See COPYING files that comes with this distribution
   8 ********************************************************************************/
   9
  10 function utf8entities($source)
  11 {
  12    // array used to figure what number to decrement from character order value
  13    // according to number of characters used to map unicode to ascii by utf-8
  14    $decrement[4] = 240;
  15    $decrement[3] = 224;
  16    $decrement[2] = 192;
  17    $decrement[1] = 0;
  18
  19    // the number of bits to shift each charNum by
  20    $shift[1][0] = 0;
  21    $shift[2][0] = 6;
  22    $shift[2][1] = 0;
  23    $shift[3][0] = 12;
  24    $shift[3][1] = 6;
  25    $shift[3][2] = 0;
  26    $shift[4][0] = 18;
  27    $shift[4][1] = 12;
  28    $shift[4][2] = 6;
  29    $shift[4][3] = 0;
  30
  31    $pos = 0;
  32    $len = strlen($source);
  33    $encodedString = '';
  34    while ($pos < $len)
  35    {
  36       $charPos = $source{$pos};
  37       $asciiPos = ord($charPos);
  38       if ($asciiPos < 128)
  39       {
  40          $encodedString .= $charPos;
  41          $pos++;
  42          continue;
  43       }
  44
  45       $i=1;
  46       if (($asciiPos >= 240) && ($asciiPos <= 255)) // 4 chars representing one unicode character
  47          $i=4;
  48       else if (($asciiPos >= 224) && ($asciiPos <= 239)) // 3 chars representing one unicode character
  49          $i=3;
  50       else if (($asciiPos >= 192) && ($asciiPos <= 223)) // 2 chars representing one unicode character
  51          $i=2;
  52       else // 1 char (lower ascii)
  53          $i=1;
  54       $thisLetter = substr($source, $pos, $i);
  55       $pos += $i;
  56
  57       // process the string representing the letter to a unicode entity
  58       $thisLen = strlen($thisLetter);
  59       $thisPos = 0;
  60       $decimalCode = 0;
  61       while ($thisPos < $thisLen)
  62       {
  63          $thisCharOrd = ord(substr($thisLetter, $thisPos, 1));
  64          if ($thisPos == 0)
  65          {
  66             $charNum = intval($thisCharOrd - $decrement[$thisLen]);
  67             $decimalCode += ($charNum << $shift[$thisLen][$thisPos]);
  68          }
  69          else
  70          {
  71             $charNum = intval($thisCharOrd - 128);
  72             $decimalCode += ($charNum << $shift[$thisLen][$thisPos]);
  73          }
  74
  75          $thisPos++;
  76       }
  77
  78       $encodedLetter = '&#'. str_pad($decimalCode, ($thisLen==1)?3:5, '0', STR_PAD_LEFT).';';
  79       $encodedString .= $encodedLetter;
  80    }
  81
  82    return $encodedString;
  83 }
  84
  85 ?>