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