banana/text.func.inc.php

   1 <?php
   2 /********************************************************************************
   3 * banana/text.php : text tools
   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 _b_($str)
  11 {
  12     if (!is_utf8($str)) {
  13         $str = utf8_encode($str);
  14     }
  15     return dgettext('banana', $str);
  16 }
  17
  18 if (!function_exists('is_utf8')) {
  19     function is_utf8($s)
  20     {
  21         return @iconv('utf-8', 'utf-8', $s) == $s;
  22     }
  23 }
  24
  25 function banana_utf8entities($source)
  26 {
  27    // array used to figure what number to decrement from character order value
  28    // according to number of characters used to map unicode to ascii by utf-8
  29    $decrement[4] = 240;
  30    $decrement[3] = 224;
  31    $decrement[2] = 192;
  32    $decrement[1] = 0;
  33
  34    // the number of bits to shift each charNum by
  35    $shift[1][0] = 0;
  36    $shift[2][0] = 6;
  37    $shift[2][1] = 0;
  38    $shift[3][0] = 12;
  39    $shift[3][1] = 6;
  40    $shift[3][2] = 0;
  41    $shift[4][0] = 18;
  42    $shift[4][1] = 12;
  43    $shift[4][2] = 6;
  44    $shift[4][3] = 0;
  45
  46    $pos = 0;
  47    $len = strlen($source);
  48    $encodedString = '';
  49    while ($pos < $len)
  50    {
  51       $charPos = $source{$pos};
  52       $asciiPos = ord($charPos);
  53       if ($asciiPos < 128)
  54       {
  55          $encodedString .= $charPos;
  56          $pos++;
  57          continue;
  58       }
  59
  60       $i=1;
  61       if (($asciiPos >= 240) && ($asciiPos <= 255)) // 4 chars representing one unicode character
  62          $i=4;
  63       else if (($asciiPos >= 224) && ($asciiPos <= 239)) // 3 chars representing one unicode character
  64          $i=3;
  65       else if (($asciiPos >= 192) && ($asciiPos <= 223)) // 2 chars representing one unicode character
  66          $i=2;
  67       else // 1 char (lower ascii)
  68          $i=1;
  69       $thisLetter = substr($source, $pos, $i);
  70       $pos += $i;
  71
  72       // process the string representing the letter to a unicode entity
  73       $thisLen = strlen($thisLetter);
  74       $thisPos = 0;
  75       $decimalCode = 0;
  76       while ($thisPos < $thisLen)
  77       {
  78          $thisCharOrd = ord(substr($thisLetter, $thisPos, 1));
  79          if ($thisPos == 0)
  80          {
  81             $charNum = intval($thisCharOrd - $decrement[$thisLen]);
  82             $decimalCode += ($charNum << $shift[$thisLen][$thisPos]);
  83          }
  84          else
  85          {
  86             $charNum = intval($thisCharOrd - 128);
  87             $decimalCode += ($charNum << $shift[$thisLen][$thisPos]);
  88          }
  89
  90          $thisPos++;
  91       }
  92
  93       $encodedLetter = '&#'. str_pad($decimalCode, ($thisLen==1)?3:5, '0', STR_PAD_LEFT).';';
  94       $encodedString .= $encodedLetter;
  95    }
  96
  97    return $encodedString;
  98 }
  99
 100 // vim:set et sw=4 sts=4 ts=4
 101 ?>