[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3  * @version $Id: unicode.php,v 1.2 2006/02/26 13:20:44 harryf Exp $
   4  * Tools for conversion between UTF-8 and unicode
   5  * The Original Code is Mozilla Communicator client code.
   6  * The Initial Developer of the Original Code is
   7  * Netscape Communications Corporation.
   8  * Portions created by the Initial Developer are Copyright (C) 1998
   9  * the Initial Developer. All Rights Reserved.
  10  * Ported to PHP by Henri Sivonen (http://hsivonen.iki.fi)
  11  * Slight modifications to fit with phputf8 library by Harry Fuecks (hfuecks gmail com)
  12  * @see http://lxr.mozilla.org/seamonkey/source/intl/uconv/src/nsUTF8ToUnicode.cpp
  13  * @see http://lxr.mozilla.org/seamonkey/source/intl/uconv/src/nsUnicodeToUTF8.cpp
  14  * @see http://hsivonen.iki.fi/php-utf8/
  15  * @package utf8
  16  * @subpackage unicode
  17  */
  18  
  19  //--------------------------------------------------------------------
  20  /**
  21  * Takes an UTF-8 string and returns an array of ints representing the 
  22  * Unicode characters. Astral planes are supported ie. the ints in the
  23  * output can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates
  24  * are not allowed.
  25  * Returns false if the input string isn't a valid UTF-8 octet sequence
  26  * and raises a PHP error at level E_USER_WARNING
  27  * Note: this function has been modified slightly in this library to
  28  * trigger errors on encountering bad bytes
  29  * @author <hsivonen@iki.fi>
  30  * @param string UTF-8 encoded string
  31  * @return mixed array of unicode code points or FALSE if UTF-8 invalid
  32  * @see utf8_from_unicode
  33  * @see http://hsivonen.iki.fi/php-utf8/
  34  * @package utf8
  35  * @subpackage unicode
  36  */
  37  function utf8_to_unicode($str) {
  38      $mState = 0;     // cached expected number of octets after the current octet
  39                       // until the beginning of the next UTF8 character sequence
  40      $mUcs4  = 0;     // cached Unicode character
  41      $mBytes = 1;     // cached expected number of octets in the current sequence
  42      
  43      $out = array();
  44      
  45      $len = strlen($str);
  46      
  47      for($i = 0; $i < $len; $i++) {
  48          
  49          $in = ord($str{$i});
  50          
  51          if ( $mState == 0) {
  52              
  53              // When mState is zero we expect either a US-ASCII character or a
  54              // multi-octet sequence.
  55              if (0 == (0x80 & ($in))) {
  56                  // US-ASCII, pass straight through.
  57                  $out[] = $in;
  58                  $mBytes = 1;
  59                  
  60              } else if (0xC0 == (0xE0 & ($in))) {
  61                  // First octet of 2 octet sequence
  62                  $mUcs4 = ($in);
  63                  $mUcs4 = ($mUcs4 & 0x1F) << 6;
  64                  $mState = 1;
  65                  $mBytes = 2;
  66                  
  67              } else if (0xE0 == (0xF0 & ($in))) {
  68                  // First octet of 3 octet sequence
  69                  $mUcs4 = ($in);
  70                  $mUcs4 = ($mUcs4 & 0x0F) << 12;
  71                  $mState = 2;
  72                  $mBytes = 3;
  73                  
  74              } else if (0xF0 == (0xF8 & ($in))) {
  75                  // First octet of 4 octet sequence
  76                  $mUcs4 = ($in);
  77                  $mUcs4 = ($mUcs4 & 0x07) << 18;
  78                  $mState = 3;
  79                  $mBytes = 4;
  80                  
  81              } else if (0xF8 == (0xFC & ($in))) {
  82                  /* First octet of 5 octet sequence.
  83                  *
  84                  * This is illegal because the encoded codepoint must be either
  85                  * (a) not the shortest form or
  86                  * (b) outside the Unicode range of 0-0x10FFFF.
  87                  * Rather than trying to resynchronize, we will carry on until the end
  88                  * of the sequence and let the later error handling code catch it.
  89                  */
  90                  $mUcs4 = ($in);
  91                  $mUcs4 = ($mUcs4 & 0x03) << 24;
  92                  $mState = 4;
  93                  $mBytes = 5;
  94                  
  95              } else if (0xFC == (0xFE & ($in))) {
  96                  // First octet of 6 octet sequence, see comments for 5 octet sequence.
  97                  $mUcs4 = ($in);
  98                  $mUcs4 = ($mUcs4 & 1) << 30;
  99                  $mState = 5;
 100                  $mBytes = 6;
 101                  
 102              } else {
 103                  /* Current octet is neither in the US-ASCII range nor a legal first
 104                   * octet of a multi-octet sequence.
 105                   */
 106                  trigger_error(
 107                          'utf8_to_unicode: Illegal sequence identifier '.
 108                              'in UTF-8 at byte '.$i,
 109                          E_USER_WARNING
 110                      );
 111                  return FALSE;
 112                  
 113              }
 114          
 115          } else {
 116              
 117              // When mState is non-zero, we expect a continuation of the multi-octet
 118              // sequence
 119              if (0x80 == (0xC0 & ($in))) {
 120                  
 121                  // Legal continuation.
 122                  $shift = ($mState - 1) * 6;
 123                  $tmp = $in;
 124                  $tmp = ($tmp & 0x0000003F) << $shift;
 125                  $mUcs4 |= $tmp;
 126              
 127                  /**
 128                  * End of the multi-octet sequence. mUcs4 now contains the final
 129                  * Unicode codepoint to be output
 130                  */
 131                  if (0 == --$mState) {
 132                      
 133                      /*
 134                      * Check for illegal sequences and codepoints.
 135                      */
 136                      // From Unicode 3.1, non-shortest form is illegal
 137                      if (((2 == $mBytes) && ($mUcs4 < 0x0080)) ||
 138                          ((3 == $mBytes) && ($mUcs4 < 0x0800)) ||
 139                          ((4 == $mBytes) && ($mUcs4 < 0x10000)) ||
 140                          (4 < $mBytes) ||
 141                          // From Unicode 3.2, surrogate characters are illegal
 142                          (($mUcs4 & 0xFFFFF800) == 0xD800) ||
 143                          // Codepoints outside the Unicode range are illegal
 144                          ($mUcs4 > 0x10FFFF)) {
 145                          
 146                          trigger_error(
 147                                  'utf8_to_unicode: Illegal sequence or codepoint '.
 148                                      'in UTF-8 at byte '.$i,
 149                                  E_USER_WARNING
 150                              );
 151                          
 152                          return FALSE;
 153                          
 154                      }
 155                      
 156                      if (0xFEFF != $mUcs4) {
 157                          // BOM is legal but we don't want to output it
 158                          $out[] = $mUcs4;
 159                      }
 160                      
 161                      //initialize UTF8 cache
 162                      $mState = 0;
 163                      $mUcs4  = 0;
 164                      $mBytes = 1;
 165                  }
 166              
 167              } else {
 168                  /**
 169                  *((0xC0 & (*in) != 0x80) && (mState != 0))
 170                  * Incomplete multi-octet sequence.
 171                  */
 172                  trigger_error(
 173                          'utf8_to_unicode: Incomplete multi-octet '.
 174                          '   sequence in UTF-8 at byte '.$i,
 175                          E_USER_WARNING
 176                      );
 177                  
 178                  return FALSE;
 179              }
 180          }
 181      }
 182      return $out;
 183  }
 184  
 185  //--------------------------------------------------------------------
 186  /**
 187  * Takes an array of ints representing the Unicode characters and returns 
 188  * a UTF-8 string. Astral planes are supported ie. the ints in the
 189  * input can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates
 190  * are not allowed.
 191  * Returns false if the input array contains ints that represent 
 192  * surrogates or are outside the Unicode range
 193  * and raises a PHP error at level E_USER_WARNING
 194  * Note: this function has been modified slightly in this library to use
 195  * output buffering to concatenate the UTF-8 string (faster) as well as
 196  * reference the array by it's keys
 197  * @param array of unicode code points representing a string
 198  * @return mixed UTF-8 string or FALSE if array contains invalid code points
 199  * @author <hsivonen@iki.fi>
 200  * @see utf8_to_unicode
 201  * @see http://hsivonen.iki.fi/php-utf8/
 202  * @package utf8
 203  * @subpackage unicode
 204  */
 205  function utf8_from_unicode($arr) {
 206      ob_start();
 207      
 208      foreach (array_keys($arr) as $k) {
 209          
 210          # ASCII range (including control chars)
 211          if ( ($arr[$k] >= 0) && ($arr[$k] <= 0x007f) ) {
 212              
 213              echo chr($arr[$k]);
 214          
 215          # 2 byte sequence
 216          } else if ($arr[$k] <= 0x07ff) {
 217              
 218              echo chr(0xc0 | ($arr[$k] >> 6));
 219              echo chr(0x80 | ($arr[$k] & 0x003f));
 220          
 221          # Byte order mark (skip)
 222          } else if($arr[$k] == 0xFEFF) {
 223              
 224              // nop -- zap the BOM
 225          
 226          # Test for illegal surrogates
 227          } else if ($arr[$k] >= 0xD800 && $arr[$k] <= 0xDFFF) {
 228              
 229              // found a surrogate
 230              trigger_error(
 231                  'utf8_from_unicode: Illegal surrogate '.
 232                      'at index: '.$k.', value: '.$arr[$k],
 233                  E_USER_WARNING
 234                  );
 235              
 236              return FALSE;
 237          
 238          # 3 byte sequence
 239          } else if ($arr[$k] <= 0xffff) {
 240              
 241              echo chr(0xe0 | ($arr[$k] >> 12));
 242              echo chr(0x80 | (($arr[$k] >> 6) & 0x003f));
 243              echo chr(0x80 | ($arr[$k] & 0x003f));
 244          
 245          # 4 byte sequence
 246          } else if ($arr[$k] <= 0x10ffff) {
 247              
 248              echo chr(0xf0 | ($arr[$k] >> 18));
 249              echo chr(0x80 | (($arr[$k] >> 12) & 0x3f));
 250              echo chr(0x80 | (($arr[$k] >> 6) & 0x3f));
 251              echo chr(0x80 | ($arr[$k] & 0x3f));
 252              
 253          } else {
 254              
 255              trigger_error(
 256                  'utf8_from_unicode: Codepoint out of Unicode range '.
 257                      'at index: '.$k.', value: '.$arr[$k],
 258                  E_USER_WARNING
 259                  );
 260              
 261              // out of range
 262              return FALSE;
 263          }
 264      }
 265      
 266      $result = ob_get_contents();
 267      ob_end_clean();
 268      return $result;
 269  }