libs\TexyUtf.php

  1: <?php
  2: 
  3: /**
  4:  * Texy! - human-readable text to HTML converter.
  5:  *
  6:  * @copyright  Copyright (c) 2004, 2010 David Grudl
  7:  * @license    GNU GENERAL PUBLIC LICENSE version 2 or 3
  8:  * @link       http://texy.info
  9:  * @package    Texy
 10:  */
 11: 
 12: 
 13: 
 14: /**
 15:  * UTF-8 helper.
 16:  *
 17:  * @copyright  Copyright (c) 2004, 2010 David Grudl
 18:  * @package    Texy
 19:  */
 20: class TexyUtf
 21: {
 22:     /** @var array */
 23:     private static $xlat;
 24: 
 25:     /** @var array */
 26:     private static $xlatCache;
 27: 
 28: 
 29: 
 30:     /**
 31:      * static class.
 32:      */
 33:     final public function __construct()
 34:     {
 35:         throw new LogicException("Cannot instantiate static class " . get_class($this));
 36:     }
 37: 
 38: 
 39: 
 40:     /**
 41:      * Converts from source encoding to UTF-8.
 42:      */
 43:     public static function toUtf($s, $encoding)
 44:     {
 45:         return iconv($encoding, 'UTF-8', $s);
 46:     }
 47: 
 48: 
 49: 
 50:     /**
 51:      * Converts from UTF-8 to dest encoding.
 52:      */
 53:     public static function utfTo($s, $encoding)
 54:     {
 55:         return iconv('utf-8', $encoding.'//TRANSLIT', $s);
 56:     }
 57: 
 58: 
 59: 
 60:     /**
 61:      * StrToLower in UTF-8.
 62:      */
 63:     public static function strtolower($s)
 64:     {
 65:         if (function_exists('mb_strtolower'))
 66:             return mb_strtolower($s, 'UTF-8');
 67: 
 68:         return @iconv('WINDOWS-1250', 'UTF-8', strtr( // intentionally @
 69:             iconv('UTF-8', 'WINDOWS-1250//IGNORE', $s),
 70:             "ABCDEFGHIJKLMNOPQRSTUVWXYZ\x8a\x8c\x8d\x8e\x8f\xa3\xa5\xaa\xaf\xbc\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2\xd3\xd4\xd5\xd6\xd8\xd9\xda\xdb\xdc\xdd\xde",
 71:             "abcdefghijklmnopqrstuvwxyz\x9a\x9c\x9d\x9e\x9f\xb3\xb9\xba\xbf\xbe\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf8\xf9\xfa\xfb\xfc\xfd\xfe"
 72:         ));
 73:     }
 74: 
 75: 
 76: 
 77:     /**
 78:      * Converts UTF-8 to ASCII.
 79:      * iconv('UTF-8', 'ASCII//TRANSLIT', ...) has problem with glibc!
 80:      */
 81:     public static function utf2ascii($s)
 82:     {
 83:         $s = strtr($s, '`\'"^~', '-----');
 84:         if (ICONV_IMPL === 'glibc') {
 85:             $s = @iconv('UTF-8', 'WINDOWS-1250//TRANSLIT', $s); // intentionally @
 86:             $s = strtr($s, "\xa5\xa3\xbc\x8c\xa7\x8a\xaa\x8d\x8f\x8e\xaf\xb9\xb3\xbe\x9c\x9a\xba\x9d\x9f\x9e\xbf\xc0\xc1\xc2\xc3\xc4\xc5\xc6\xc7\xc8\xc9\xca\xcb\xcc\xcd\xce\xcf\xd0\xd1\xd2"
 87:                 ."\xd3\xd4\xd5\xd6\xd7\xd8\xd9\xda\xdb\xdc\xdd\xde\xdf\xe0\xe1\xe2\xe3\xe4\xe5\xe6\xe7\xe8\xe9\xea\xeb\xec\xed\xee\xef\xf0\xf1\xf2\xf3\xf4\xf5\xf6\xf8\xf9\xfa\xfb\xfc\xfd\xfe",
 88:                 "ALLSSSSTZZZallssstzzzRAAAALCCCEEEEIIDDNNOOOOxRUUUUYTsraaaalccceeeeiiddnnooooruuuuyt");
 89:         } else {
 90:             $s = @iconv('UTF-8', 'ASCII//TRANSLIT', $s); // intentionally @
 91:         }
 92:         $s = str_replace(array('`', "'", '"', '^', '~'), '', $s);
 93:         return $s;
 94:     }
 95: 
 96: 
 97: 
 98:     /**
 99:      * Converts UTF-8 to dest encoding + html entities.
100:      */
101:     public static function utf2html($s, $encoding)
102:     {
103:         // convert from UTF-8
104:         if (strcasecmp($encoding, 'utf-8') === 0) return $s;
105: 
106:         // prepare UTF-8 -> charset table
107:         self::$xlat = & self::$xlatCache[strtolower($encoding)];
108:         if (!self::$xlat) {
109:             for ($i = 128; $i<256; $i++) {
110:                 $ch = @iconv($encoding, 'UTF-8//IGNORE', chr($i)); // intentionally @
111:                 if ($ch) self::$xlat[$ch] = chr($i);
112:             }
113:         }
114: 
115:         // convert
116:         return preg_replace_callback('#[\x80-\x{FFFF}]#u', array(__CLASS__, 'cb'), $s);
117:     }
118: 
119: 
120: 
121:     /**
122:      * Callback; converts UTF-8 to HTML entity OR character in dest encoding.
123:      */
124:     private static function cb($m)
125:     {
126:         $m = $m[0];
127:         if (isset(self::$xlat[$m])) return self::$xlat[$m];
128: 
129:         $ch1 = ord($m[0]);
130:         $ch2 = ord($m[1]);
131:         if (($ch2 >> 6) !== 2) return '';
132: 
133:         if (($ch1 & 0xE0) === 0xC0)
134:             return '&#' . ((($ch1 & 0x1F) << 6) + ($ch2 & 0x3F)) . ';';
135: 
136:         if (($ch1 & 0xF0) === 0xE0) {
137:             $ch3 = ord($m[2]);
138:             if (($ch3 >> 6) !== 2) return '';
139:             return '&#' . ((($ch1 & 0xF) << 12) + (($ch2 & 0x3F) << 06) + (($ch3 & 0x3F))) . ';';
140:         }
141: 
142:         return '';
143:     }
144: 
145: }
146:
Namespaces

Classes

Interfaces
