[Summary view] [Print] [Text view]

   1  <?php
   2  /**
   3  * Tools for conversion between UTF-8 and unicode
   4  * The Original Code is Mozilla Communicator client code.
   5  * The Initial Developer of the Original Code is
   6  * Netscape Communications Corporation.
   7  * Portions created by the Initial Developer are Copyright (C) 1998
   8  * the Initial Developer. All Rights Reserved.
   9  * Ported to PHP by Henri Sivonen (http://hsivonen.iki.fi)
  10  * Slight modifications to fit with phputf8 library by Harry Fuecks (hfuecks gmail com)
  11  * @see http://lxr.mozilla.org/seamonkey/source/intl/uconv/src/nsUTF8ToUnicode.cpp
  12  * @see http://lxr.mozilla.org/seamonkey/source/intl/uconv/src/nsUnicodeToUTF8.cpp
  13  * @see http://hsivonen.iki.fi/php-utf8/
  14  * @package utf8
  15  */
  16  
  17  //--------------------------------------------------------------------
  18  /**
  19  * Takes an UTF-8 string and returns an array of ints representing the
  20  * Unicode characters. Astral planes are supported ie. the ints in the
  21  * output can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates
  22  * are not allowed.
  23  * Returns false if the input string isn't a valid UTF-8 octet sequence
  24  * and raises a PHP error at level E_USER_WARNING
  25  * Note: this function has been modified slightly in this library to
  26  * trigger errors on encountering bad bytes
  27  *
  28  * Joomla modification - As of PHP 7.4, curly brace access has been deprecated. As a result this function has been
  29  * modified to use square brace syntax
  30  * See https://github.com/php/php-src/commit/d574df63dc375f5fc9202ce5afde23f866b6450a
  31  * for additional references
  32  *
  33  * @author <[email protected]>
  34  * @param string UTF-8 encoded string
  35  * @return mixed array of unicode code points or FALSE if UTF-8 invalid
  36  * @see utf8_from_unicode
  37  * @see http://hsivonen.iki.fi/php-utf8/
  38  * @package utf8
  39  */
  40  function utf8_to_unicode($str) {
  41      $mState = 0;     // cached expected number of octets after the current octet
  42                       // until the beginning of the next UTF8 character sequence
  43      $mUcs4  = 0;     // cached Unicode character
  44      $mBytes = 1;     // cached expected number of octets in the current sequence
  45  
  46      $out = array();
  47  
  48      $len = strlen($str);
  49  
  50      for($i = 0; $i < $len; $i++) {
  51  
  52          $in = ord($str[$i]);
  53  
  54          if ( $mState == 0) {
  55  
  56              // When mState is zero we expect either a US-ASCII character or a
  57              // multi-octet sequence.
  58              if (0 == (0x80 & ($in))) {
  59                  // US-ASCII, pass straight through.
  60                  $out[] = $in;
  61                  $mBytes = 1;
  62  
  63              } else if (0xC0 == (0xE0 & ($in))) {
  64                  // First octet of 2 octet sequence
  65                  $mUcs4 = ($in);
  66                  $mUcs4 = ($mUcs4 & 0x1F) << 6;
  67                  $mState = 1;
  68                  $mBytes = 2;
  69  
  70              } else if (0xE0 == (0xF0 & ($in))) {
  71                  // First octet of 3 octet sequence
  72                  $mUcs4 = ($in);
  73                  $mUcs4 = ($mUcs4 & 0x0F) << 12;
  74                  $mState = 2;
  75                  $mBytes = 3;
  76  
  77              } else if (0xF0 == (0xF8 & ($in))) {
  78                  // First octet of 4 octet sequence
  79                  $mUcs4 = ($in);
  80                  $mUcs4 = ($mUcs4 & 0x07) << 18;
  81                  $mState = 3;
  82                  $mBytes = 4;
  83  
  84              } else if (0xF8 == (0xFC & ($in))) {
  85                  /* First octet of 5 octet sequence.
  86                  *
  87                  * This is illegal because the encoded codepoint must be either
  88                  * (a) not the shortest form or
  89                  * (b) outside the Unicode range of 0-0x10FFFF.
  90                  * Rather than trying to resynchronize, we will carry on until the end
  91                  * of the sequence and let the later error handling code catch it.
  92                  */
  93                  $mUcs4 = ($in);
  94                  $mUcs4 = ($mUcs4 & 0x03) << 24;
  95                  $mState = 4;
  96                  $mBytes = 5;
  97  
  98              } else if (0xFC == (0xFE & ($in))) {
  99                  // First octet of 6 octet sequence, see comments for 5 octet sequence.
 100                  $mUcs4 = ($in);
 101                  $mUcs4 = ($mUcs4 & 1) << 30;
 102                  $mState = 5;
 103                  $mBytes = 6;
 104  
 105              } else {
 106                  /* Current octet is neither in the US-ASCII range nor a legal first
 107                   * octet of a multi-octet sequence.
 108                   */
 109                  trigger_error(
 110                          'utf8_to_unicode: Illegal sequence identifier '.
 111                              'in UTF-8 at byte '.$i,
 112                          E_USER_WARNING
 113                      );
 114                  return FALSE;
 115  
 116              }
 117  
 118          } else {
 119  
 120              // When mState is non-zero, we expect a continuation of the multi-octet
 121              // sequence
 122              if (0x80 == (0xC0 & ($in))) {
 123  
 124                  // Legal continuation.
 125                  $shift = ($mState - 1) * 6;
 126                  $tmp = $in;
 127                  $tmp = ($tmp & 0x0000003F) << $shift;
 128                  $mUcs4 |= $tmp;
 129  
 130                  /**
 131                  * End of the multi-octet sequence. mUcs4 now contains the final
 132                  * Unicode codepoint to be output
 133                  */
 134                  if (0 == --$mState) {
 135  
 136                      /*
 137                      * Check for illegal sequences and codepoints.
 138                      */
 139                      // From Unicode 3.1, non-shortest form is illegal
 140                      if (((2 == $mBytes) && ($mUcs4 < 0x0080)) ||
 141                          ((3 == $mBytes) && ($mUcs4 < 0x0800)) ||
 142                          ((4 == $mBytes) && ($mUcs4 < 0x10000)) ||
 143                          (4 < $mBytes) ||
 144                          // From Unicode 3.2, surrogate characters are illegal
 145                          (($mUcs4 & 0xFFFFF800) == 0xD800) ||
 146                          // Codepoints outside the Unicode range are illegal
 147                          ($mUcs4 > 0x10FFFF)) {
 148  
 149                          trigger_error(
 150                                  'utf8_to_unicode: Illegal sequence or codepoint '.
 151                                      'in UTF-8 at byte '.$i,
 152                                  E_USER_WARNING
 153                              );
 154  
 155                          return FALSE;
 156  
 157                      }
 158  
 159                      if (0xFEFF != $mUcs4) {
 160                          // BOM is legal but we don't want to output it
 161                          $out[] = $mUcs4;
 162                      }
 163  
 164                      //initialize UTF8 cache
 165                      $mState = 0;
 166                      $mUcs4  = 0;
 167                      $mBytes = 1;
 168                  }
 169  
 170              } else {
 171                  /**
 172                  *((0xC0 & (*in) != 0x80) && (mState != 0))
 173                  * Incomplete multi-octet sequence.
 174                  */
 175                  trigger_error(
 176                          'utf8_to_unicode: Incomplete multi-octet '.
 177                          '   sequence in UTF-8 at byte '.$i,
 178                          E_USER_WARNING
 179                      );
 180  
 181                  return FALSE;
 182              }
 183          }
 184      }
 185      return $out;
 186  }
 187  
 188  //--------------------------------------------------------------------
 189  /**
 190  * Takes an array of ints representing the Unicode characters and returns
 191  * a UTF-8 string. Astral planes are supported ie. the ints in the
 192  * input can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates
 193  * are not allowed.
 194  * Returns false if the input array contains ints that represent
 195  * surrogates or are outside the Unicode range
 196  * and raises a PHP error at level E_USER_WARNING
 197  * Note: this function has been modified slightly in this library to use
 198  * output buffering to concatenate the UTF-8 string (faster) as well as
 199  * reference the array by it's keys
 200  * @param array of unicode code points representing a string
 201  * @return mixed UTF-8 string or FALSE if array contains invalid code points
 202  * @author <[email protected]>
 203  * @see utf8_to_unicode
 204  * @see http://hsivonen.iki.fi/php-utf8/
 205  * @package utf8
 206  */
 207  function utf8_from_unicode($arr) {
 208      ob_start();
 209  
 210      foreach (array_keys($arr) as $k) {
 211  
 212          # ASCII range (including control chars)
 213          if ( ($arr[$k] >= 0) && ($arr[$k] <= 0x007f) ) {
 214  
 215              echo chr($arr[$k]);
 216  
 217          # 2 byte sequence
 218          } else if ($arr[$k] <= 0x07ff) {
 219  
 220              echo chr(0xc0 | ($arr[$k] >> 6));
 221              echo chr(0x80 | ($arr[$k] & 0x003f));
 222  
 223          # Byte order mark (skip)
 224          } else if($arr[$k] == 0xFEFF) {
 225  
 226              // nop -- zap the BOM
 227  
 228          # Test for illegal surrogates
 229          } else if ($arr[$k] >= 0xD800 && $arr[$k] <= 0xDFFF) {
 230  
 231              // found a surrogate
 232              trigger_error(
 233                  'utf8_from_unicode: Illegal surrogate '.
 234                      'at index: '.$k.', value: '.$arr[$k],
 235                  E_USER_WARNING
 236                  );
 237  
 238              return FALSE;
 239  
 240          # 3 byte sequence
 241          } else if ($arr[$k] <= 0xffff) {
 242  
 243              echo chr(0xe0 | ($arr[$k] >> 12));
 244              echo chr(0x80 | (($arr[$k] >> 6) & 0x003f));
 245              echo chr(0x80 | ($arr[$k] & 0x003f));
 246  
 247          # 4 byte sequence
 248          } else if ($arr[$k] <= 0x10ffff) {
 249  
 250              echo chr(0xf0 | ($arr[$k] >> 18));
 251              echo chr(0x80 | (($arr[$k] >> 12) & 0x3f));
 252              echo chr(0x80 | (($arr[$k] >> 6) & 0x3f));
 253              echo chr(0x80 | ($arr[$k] & 0x3f));
 254  
 255          } else {
 256  
 257              trigger_error(
 258                  'utf8_from_unicode: Codepoint out of Unicode range '.
 259                      'at index: '.$k.', value: '.$arr[$k],
 260                  E_USER_WARNING
 261                  );
 262  
 263              // out of range
 264              return FALSE;
 265          }
 266      }
 267  
 268      $result = ob_get_contents();
 269      ob_end_clean();
 270      return $result;
 271  }