1<?php 2 3namespace dokuwiki\Utf8; 4 5/** 6 * Convert between UTF-8 and a list of Unicode Code Points 7 */ 8class Unicode 9{ 10 11 /** 12 * Takes an UTF-8 string and returns an array of ints representing the 13 * Unicode characters. Astral planes are supported ie. the ints in the 14 * output can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates 15 * are not allowed. 16 * 17 * If $strict is set to true the function returns false if the input 18 * string isn't a valid UTF-8 octet sequence and raises a PHP error at 19 * level E_USER_WARNING 20 * 21 * Note: this function has been modified slightly in this library to 22 * trigger errors on encountering bad bytes 23 * 24 * @author <hsivonen@iki.fi> 25 * @author Harry Fuecks <hfuecks@gmail.com> 26 * @see unicode_to_utf8 27 * @link http://hsivonen.iki.fi/php-utf8/ 28 * @link http://sourceforge.net/projects/phputf8/ 29 * @todo break into less complex chunks 30 * @todo use exceptions instead of user errors 31 * 32 * @param string $str UTF-8 encoded string 33 * @param boolean $strict Check for invalid sequences? 34 * @return mixed array of unicode code points or false if UTF-8 invalid 35 */ 36 public static function fromUtf8($str, $strict = false) 37 { 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 = []; 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 // When mState is zero we expect either a US-ASCII character or a 53 // multi-octet sequence. 54 if (0 === (0x80 & $in)) { 55 // US-ASCII, pass straight through. 56 $out[] = $in; 57 $mBytes = 1; 58 } elseif (0xC0 === (0xE0 & $in)) { 59 // First octet of 2 octet sequence 60 $mUcs4 = $in; 61 $mUcs4 = ($mUcs4 & 0x1F) << 6; 62 $mState = 1; 63 $mBytes = 2; 64 } elseif (0xE0 === (0xF0 & $in)) { 65 // First octet of 3 octet sequence 66 $mUcs4 = $in; 67 $mUcs4 = ($mUcs4 & 0x0F) << 12; 68 $mState = 2; 69 $mBytes = 3; 70 } elseif (0xF0 === (0xF8 & $in)) { 71 // First octet of 4 octet sequence 72 $mUcs4 = $in; 73 $mUcs4 = ($mUcs4 & 0x07) << 18; 74 $mState = 3; 75 $mBytes = 4; 76 } elseif (0xF8 === (0xFC & $in)) { 77 /* First octet of 5 octet sequence. 78 * 79 * This is illegal because the encoded codepoint must be either 80 * (a) not the shortest form or 81 * (b) outside the Unicode range of 0-0x10FFFF. 82 * Rather than trying to resynchronize, we will carry on until the end 83 * of the sequence and let the later error handling code catch it. 84 */ 85 $mUcs4 = $in; 86 $mUcs4 = ($mUcs4 & 0x03) << 24; 87 $mState = 4; 88 $mBytes = 5; 89 } elseif (0xFC === (0xFE & $in)) { 90 // First octet of 6 octet sequence, see comments for 5 octet sequence. 91 $mUcs4 = $in; 92 $mUcs4 = ($mUcs4 & 1) << 30; 93 $mState = 5; 94 $mBytes = 6; 95 } elseif ($strict) { 96 /* Current octet is neither in the US-ASCII range nor a legal first 97 * octet of a multi-octet sequence. 98 */ 99 trigger_error( 100 'utf8_to_unicode: Illegal sequence identifier ' . 101 'in UTF-8 at byte ' . $i, 102 E_USER_WARNING 103 ); 104 return false; 105 106 } 107 } elseif (0x80 === (0xC0 & $in)) { 108 // When mState is non-zero, we expect a continuation of the multi-octet 109 // sequence 110 // Legal continuation. 111 $shift = ($mState - 1) * 6; 112 $tmp = $in; 113 $tmp = ($tmp & 0x0000003F) << $shift; 114 $mUcs4 |= $tmp; 115 /** 116 * End of the multi-octet sequence. mUcs4 now contains the final 117 * Unicode codepoint to be output 118 */ 119 if (0 === --$mState) { 120 121 /* 122 * Check for illegal sequences and codepoints. 123 */ 124 // From Unicode 3.1, non-shortest form is illegal 125 if (((2 === $mBytes) && ($mUcs4 < 0x0080)) || 126 ((3 === $mBytes) && ($mUcs4 < 0x0800)) || 127 ((4 === $mBytes) && ($mUcs4 < 0x10000)) || 128 (4 < $mBytes) || 129 // From Unicode 3.2, surrogate characters are illegal 130 (($mUcs4 & 0xFFFFF800) === 0xD800) || 131 // Codepoints outside the Unicode range are illegal 132 ($mUcs4 > 0x10FFFF)) { 133 134 if ($strict) { 135 trigger_error( 136 'utf8_to_unicode: Illegal sequence or codepoint ' . 137 'in UTF-8 at byte ' . $i, 138 E_USER_WARNING 139 ); 140 141 return false; 142 } 143 144 } 145 146 if (0xFEFF !== $mUcs4) { 147 // BOM is legal but we don't want to output it 148 $out[] = $mUcs4; 149 } 150 151 //initialize UTF8 cache 152 $mState = 0; 153 $mUcs4 = 0; 154 $mBytes = 1; 155 } 156 } elseif ($strict) { 157 /** 158 *((0xC0 & (*in) != 0x80) && (mState != 0)) 159 * Incomplete multi-octet sequence. 160 */ 161 trigger_error( 162 'utf8_to_unicode: Incomplete multi-octet ' . 163 ' sequence in UTF-8 at byte ' . $i, 164 E_USER_WARNING 165 ); 166 167 return false; 168 } 169 } 170 return $out; 171 } 172 173 /** 174 * Takes an array of ints representing the Unicode characters and returns 175 * a UTF-8 string. Astral planes are supported ie. the ints in the 176 * input can be > 0xFFFF. Occurrances of the BOM are ignored. Surrogates 177 * are not allowed. 178 * 179 * If $strict is set to true the function returns false if the input 180 * array contains ints that represent surrogates or are outside the 181 * Unicode range and raises a PHP error at level E_USER_WARNING 182 * 183 * Note: this function has been modified slightly in this library to use 184 * output buffering to concatenate the UTF-8 string (faster) as well as 185 * reference the array by it's keys 186 * 187 * @param array $arr of unicode code points representing a string 188 * @param boolean $strict Check for invalid sequences? 189 * @return string|false UTF-8 string or false if array contains invalid code points 190 * 191 * @author <hsivonen@iki.fi> 192 * @author Harry Fuecks <hfuecks@gmail.com> 193 * @see utf8_to_unicode 194 * @link http://hsivonen.iki.fi/php-utf8/ 195 * @link http://sourceforge.net/projects/phputf8/ 196 * @todo use exceptions instead of user errors 197 */ 198 public static function toUtf8($arr, $strict = false) 199 { 200 if (!is_array($arr)) return ''; 201 ob_start(); 202 203 foreach (array_keys($arr) as $k) { 204 205 if (($arr[$k] >= 0) && ($arr[$k] <= 0x007f)) { 206 # ASCII range (including control chars) 207 echo chr($arr[$k]); 208 } elseif ($arr[$k] <= 0x07ff) { 209 # 2 byte sequence 210 echo chr(0xc0 | ($arr[$k] >> 6)); 211 echo chr(0x80 | ($arr[$k] & 0x003f)); 212 } elseif ($arr[$k] == 0xFEFF) { 213 # Byte order mark (skip) 214 // nop -- zap the BOM 215 } elseif ($arr[$k] >= 0xD800 && $arr[$k] <= 0xDFFF) { 216 # Test for illegal surrogates 217 // found a surrogate 218 if ($strict) { 219 trigger_error( 220 'unicode_to_utf8: Illegal surrogate ' . 221 'at index: ' . $k . ', value: ' . $arr[$k], 222 E_USER_WARNING 223 ); 224 return false; 225 } 226 } elseif ($arr[$k] <= 0xffff) { 227 # 3 byte sequence 228 echo chr(0xe0 | ($arr[$k] >> 12)); 229 echo chr(0x80 | (($arr[$k] >> 6) & 0x003f)); 230 echo chr(0x80 | ($arr[$k] & 0x003f)); 231 } elseif ($arr[$k] <= 0x10ffff) { 232 # 4 byte sequence 233 echo chr(0xf0 | ($arr[$k] >> 18)); 234 echo chr(0x80 | (($arr[$k] >> 12) & 0x3f)); 235 echo chr(0x80 | (($arr[$k] >> 6) & 0x3f)); 236 echo chr(0x80 | ($arr[$k] & 0x3f)); 237 } elseif ($strict) { 238 239 trigger_error( 240 'unicode_to_utf8: Codepoint out of Unicode range ' . 241 'at index: ' . $k . ', value: ' . $arr[$k], 242 E_USER_WARNING 243 ); 244 245 // out of range 246 return false; 247 } 248 } 249 250 return ob_get_clean(); 251 } 252} 253