[ Index ]

PHP Cross Reference of WordPress Trunk (Updated Daily)

Search

title

Body

[close]

/wp-includes/Requests/ -> IDNAEncoder.php (source)

   1  <?php
   2  
   3  /**
   4   * IDNA URL encoder
   5   *
   6   * Note: Not fully compliant, as nameprep does nothing yet.
   7   *
   8   * @package Requests
   9   * @subpackage Utilities
  10   * @see https://tools.ietf.org/html/rfc3490 IDNA specification
  11   * @see https://tools.ietf.org/html/rfc3492 Punycode/Bootstrap specification
  12   */
  13  class Requests_IDNAEncoder {
  14      /**
  15       * ACE prefix used for IDNA
  16       *
  17       * @see https://tools.ietf.org/html/rfc3490#section-5
  18       * @var string
  19       */
  20      const ACE_PREFIX = 'xn--';
  21  
  22      /**#@+
  23       * Bootstrap constant for Punycode
  24       *
  25       * @see https://tools.ietf.org/html/rfc3492#section-5
  26       * @var int
  27       */
  28      const BOOTSTRAP_BASE         = 36;
  29      const BOOTSTRAP_TMIN         = 1;
  30      const BOOTSTRAP_TMAX         = 26;
  31      const BOOTSTRAP_SKEW         = 38;
  32      const BOOTSTRAP_DAMP         = 700;
  33      const BOOTSTRAP_INITIAL_BIAS = 72;
  34      const BOOTSTRAP_INITIAL_N    = 128;
  35      /**#@-*/
  36  
  37      /**
  38       * Encode a hostname using Punycode
  39       *
  40       * @param string $string Hostname
  41       * @return string Punycode-encoded hostname
  42       */
  43  	public static function encode($string) {
  44          $parts = explode('.', $string);
  45          foreach ($parts as &$part) {
  46              $part = self::to_ascii($part);
  47          }
  48          return implode('.', $parts);
  49      }
  50  
  51      /**
  52       * Convert a UTF-8 string to an ASCII string using Punycode
  53       *
  54       * @throws Requests_Exception Provided string longer than 64 ASCII characters (`idna.provided_too_long`)
  55       * @throws Requests_Exception Prepared string longer than 64 ASCII characters (`idna.prepared_too_long`)
  56       * @throws Requests_Exception Provided string already begins with xn-- (`idna.provided_is_prefixed`)
  57       * @throws Requests_Exception Encoded string longer than 64 ASCII characters (`idna.encoded_too_long`)
  58       *
  59       * @param string $string ASCII or UTF-8 string (max length 64 characters)
  60       * @return string ASCII string
  61       */
  62  	public static function to_ascii($string) {
  63          // Step 1: Check if the string is already ASCII
  64          if (self::is_ascii($string)) {
  65              // Skip to step 7
  66              if (strlen($string) < 64) {
  67                  return $string;
  68              }
  69  
  70              throw new Requests_Exception('Provided string is too long', 'idna.provided_too_long', $string);
  71          }
  72  
  73          // Step 2: nameprep
  74          $string = self::nameprep($string);
  75  
  76          // Step 3: UseSTD3ASCIIRules is false, continue
  77          // Step 4: Check if it's ASCII now
  78          if (self::is_ascii($string)) {
  79              // Skip to step 7
  80              if (strlen($string) < 64) {
  81                  return $string;
  82              }
  83  
  84              throw new Requests_Exception('Prepared string is too long', 'idna.prepared_too_long', $string);
  85          }
  86  
  87          // Step 5: Check ACE prefix
  88          if (strpos($string, self::ACE_PREFIX) === 0) {
  89              throw new Requests_Exception('Provided string begins with ACE prefix', 'idna.provided_is_prefixed', $string);
  90          }
  91  
  92          // Step 6: Encode with Punycode
  93          $string = self::punycode_encode($string);
  94  
  95          // Step 7: Prepend ACE prefix
  96          $string = self::ACE_PREFIX . $string;
  97  
  98          // Step 8: Check size
  99          if (strlen($string) < 64) {
 100              return $string;
 101          }
 102  
 103          throw new Requests_Exception('Encoded string is too long', 'idna.encoded_too_long', $string);
 104      }
 105  
 106      /**
 107       * Check whether a given string contains only ASCII characters
 108       *
 109       * @internal (Testing found regex was the fastest implementation)
 110       *
 111       * @param string $string
 112       * @return bool Is the string ASCII-only?
 113       */
 114  	protected static function is_ascii($string) {
 115          return (preg_match('/(?:[^\x00-\x7F])/', $string) !== 1);
 116      }
 117  
 118      /**
 119       * Prepare a string for use as an IDNA name
 120       *
 121       * @todo Implement this based on RFC 3491 and the newer 5891
 122       * @param string $string
 123       * @return string Prepared string
 124       */
 125  	protected static function nameprep($string) {
 126          return $string;
 127      }
 128  
 129      /**
 130       * Convert a UTF-8 string to a UCS-4 codepoint array
 131       *
 132       * Based on Requests_IRI::replace_invalid_with_pct_encoding()
 133       *
 134       * @throws Requests_Exception Invalid UTF-8 codepoint (`idna.invalidcodepoint`)
 135       * @param string $input
 136       * @return array Unicode code points
 137       */
 138  	protected static function utf8_to_codepoints($input) {
 139          $codepoints = array();
 140  
 141          // Get number of bytes
 142          $strlen = strlen($input);
 143  
 144          // phpcs:ignore Generic.CodeAnalysis.JumbledIncrementer -- This is a deliberate choice.
 145          for ($position = 0; $position < $strlen; $position++) {
 146              $value = ord($input[$position]);
 147  
 148              // One byte sequence:
 149              if ((~$value & 0x80) === 0x80) {
 150                  $character = $value;
 151                  $length    = 1;
 152                  $remaining = 0;
 153              }
 154              // Two byte sequence:
 155              elseif (($value & 0xE0) === 0xC0) {
 156                  $character = ($value & 0x1F) << 6;
 157                  $length    = 2;
 158                  $remaining = 1;
 159              }
 160              // Three byte sequence:
 161              elseif (($value & 0xF0) === 0xE0) {
 162                  $character = ($value & 0x0F) << 12;
 163                  $length    = 3;
 164                  $remaining = 2;
 165              }
 166              // Four byte sequence:
 167              elseif (($value & 0xF8) === 0xF0) {
 168                  $character = ($value & 0x07) << 18;
 169                  $length    = 4;
 170                  $remaining = 3;
 171              }
 172              // Invalid byte:
 173              else {
 174                  throw new Requests_Exception('Invalid Unicode codepoint', 'idna.invalidcodepoint', $value);
 175              }
 176  
 177              if ($remaining > 0) {
 178                  if ($position + $length > $strlen) {
 179                      throw new Requests_Exception('Invalid Unicode codepoint', 'idna.invalidcodepoint', $character);
 180                  }
 181                  for ($position++; $remaining > 0; $position++) {
 182                      $value = ord($input[$position]);
 183  
 184                      // If it is invalid, count the sequence as invalid and reprocess the current byte:
 185                      if (($value & 0xC0) !== 0x80) {
 186                          throw new Requests_Exception('Invalid Unicode codepoint', 'idna.invalidcodepoint', $character);
 187                      }
 188  
 189                      --$remaining;
 190                      $character |= ($value & 0x3F) << ($remaining * 6);
 191                  }
 192                  $position--;
 193              }
 194  
 195              if (// Non-shortest form sequences are invalid
 196                  $length > 1 && $character <= 0x7F
 197                  || $length > 2 && $character <= 0x7FF
 198                  || $length > 3 && $character <= 0xFFFF
 199                  // Outside of range of ucschar codepoints
 200                  // Noncharacters
 201                  || ($character & 0xFFFE) === 0xFFFE
 202                  || $character >= 0xFDD0 && $character <= 0xFDEF
 203                  || (
 204                      // Everything else not in ucschar
 205                      $character > 0xD7FF && $character < 0xF900
 206                      || $character < 0x20
 207                      || $character > 0x7E && $character < 0xA0
 208                      || $character > 0xEFFFD
 209                  )
 210              ) {
 211                  throw new Requests_Exception('Invalid Unicode codepoint', 'idna.invalidcodepoint', $character);
 212              }
 213  
 214              $codepoints[] = $character;
 215          }
 216  
 217          return $codepoints;
 218      }
 219  
 220      /**
 221       * RFC3492-compliant encoder
 222       *
 223       * @internal Pseudo-code from Section 6.3 is commented with "#" next to relevant code
 224       * @throws Requests_Exception On character outside of the domain (never happens with Punycode) (`idna.character_outside_domain`)
 225       *
 226       * @param string $input UTF-8 encoded string to encode
 227       * @return string Punycode-encoded string
 228       */
 229  	public static function punycode_encode($input) {
 230          $output = '';
 231          // let n = initial_n
 232          $n = self::BOOTSTRAP_INITIAL_N;
 233          // let delta = 0
 234          $delta = 0;
 235          // let bias = initial_bias
 236          $bias = self::BOOTSTRAP_INITIAL_BIAS;
 237          // let h = b = the number of basic code points in the input
 238          $h = 0;
 239          $b = 0; // see loop
 240          // copy them to the output in order
 241          $codepoints = self::utf8_to_codepoints($input);
 242          $extended   = array();
 243  
 244          foreach ($codepoints as $char) {
 245              if ($char < 128) {
 246                  // Character is valid ASCII
 247                  // TODO: this should also check if it's valid for a URL
 248                  $output .= chr($char);
 249                  $h++;
 250              }
 251              // Check if the character is non-ASCII, but below initial n
 252              // This never occurs for Punycode, so ignore in coverage
 253              // @codeCoverageIgnoreStart
 254              elseif ($char < $n) {
 255                  throw new Requests_Exception('Invalid character', 'idna.character_outside_domain', $char);
 256              }
 257              // @codeCoverageIgnoreEnd
 258              else {
 259                  $extended[$char] = true;
 260              }
 261          }
 262          $extended = array_keys($extended);
 263          sort($extended);
 264          $b = $h;
 265          // [copy them] followed by a delimiter if b > 0
 266          if (strlen($output) > 0) {
 267              $output .= '-';
 268          }
 269          // {if the input contains a non-basic code point < n then fail}
 270          // while h < length(input) do begin
 271          $codepointcount = count($codepoints);
 272          while ($h < $codepointcount) {
 273              // let m = the minimum code point >= n in the input
 274              $m = array_shift($extended);
 275              //printf('next code point to insert is %s' . PHP_EOL, dechex($m));
 276              // let delta = delta + (m - n) * (h + 1), fail on overflow
 277              $delta += ($m - $n) * ($h + 1);
 278              // let n = m
 279              $n = $m;
 280              // for each code point c in the input (in order) do begin
 281              for ($num = 0; $num < $codepointcount; $num++) {
 282                  $c = $codepoints[$num];
 283                  // if c < n then increment delta, fail on overflow
 284                  if ($c < $n) {
 285                      $delta++;
 286                  }
 287                  // if c == n then begin
 288                  elseif ($c === $n) {
 289                      // let q = delta
 290                      $q = $delta;
 291                      // for k = base to infinity in steps of base do begin
 292                      for ($k = self::BOOTSTRAP_BASE; ; $k += self::BOOTSTRAP_BASE) {
 293                          // let t = tmin if k <= bias {+ tmin}, or
 294                          //     tmax if k >= bias + tmax, or k - bias otherwise
 295                          if ($k <= ($bias + self::BOOTSTRAP_TMIN)) {
 296                              $t = self::BOOTSTRAP_TMIN;
 297                          }
 298                          elseif ($k >= ($bias + self::BOOTSTRAP_TMAX)) {
 299                              $t = self::BOOTSTRAP_TMAX;
 300                          }
 301                          else {
 302                              $t = $k - $bias;
 303                          }
 304                          // if q < t then break
 305                          if ($q < $t) {
 306                              break;
 307                          }
 308                          // output the code point for digit t + ((q - t) mod (base - t))
 309                          $digit   = $t + (($q - $t) % (self::BOOTSTRAP_BASE - $t));
 310                          $output .= self::digit_to_char($digit);
 311                          // let q = (q - t) div (base - t)
 312                          $q = floor(($q - $t) / (self::BOOTSTRAP_BASE - $t));
 313                      } // end
 314                      // output the code point for digit q
 315                      $output .= self::digit_to_char($q);
 316                      // let bias = adapt(delta, h + 1, test h equals b?)
 317                      $bias = self::adapt($delta, $h + 1, $h === $b);
 318                      // let delta = 0
 319                      $delta = 0;
 320                      // increment h
 321                      $h++;
 322                  } // end
 323              } // end
 324              // increment delta and n
 325              $delta++;
 326              $n++;
 327          } // end
 328  
 329          return $output;
 330      }
 331  
 332      /**
 333       * Convert a digit to its respective character
 334       *
 335       * @see https://tools.ietf.org/html/rfc3492#section-5
 336       * @throws Requests_Exception On invalid digit (`idna.invalid_digit`)
 337       *
 338       * @param int $digit Digit in the range 0-35
 339       * @return string Single character corresponding to digit
 340       */
 341  	protected static function digit_to_char($digit) {
 342          // @codeCoverageIgnoreStart
 343          // As far as I know, this never happens, but still good to be sure.
 344          if ($digit < 0 || $digit > 35) {
 345              throw new Requests_Exception(sprintf('Invalid digit %d', $digit), 'idna.invalid_digit', $digit);
 346          }
 347          // @codeCoverageIgnoreEnd
 348          $digits = 'abcdefghijklmnopqrstuvwxyz0123456789';
 349          return substr($digits, $digit, 1);
 350      }
 351  
 352      /**
 353       * Adapt the bias
 354       *
 355       * @see https://tools.ietf.org/html/rfc3492#section-6.1
 356       * @param int $delta
 357       * @param int $numpoints
 358       * @param bool $firsttime
 359       * @return int New bias
 360       *
 361       * function adapt(delta,numpoints,firsttime):
 362       */
 363  	protected static function adapt($delta, $numpoints, $firsttime) {
 364          // if firsttime then let delta = delta div damp
 365          if ($firsttime) {
 366              $delta = floor($delta / self::BOOTSTRAP_DAMP);
 367          }
 368          // else let delta = delta div 2
 369          else {
 370              $delta = floor($delta / 2);
 371          }
 372          // let delta = delta + (delta div numpoints)
 373          $delta += floor($delta / $numpoints);
 374          // let k = 0
 375          $k = 0;
 376          // while delta > ((base - tmin) * tmax) div 2 do begin
 377          $max = floor(((self::BOOTSTRAP_BASE - self::BOOTSTRAP_TMIN) * self::BOOTSTRAP_TMAX) / 2);
 378          while ($delta > $max) {
 379              // let delta = delta div (base - tmin)
 380              $delta = floor($delta / (self::BOOTSTRAP_BASE - self::BOOTSTRAP_TMIN));
 381              // let k = k + base
 382              $k += self::BOOTSTRAP_BASE;
 383          } // end
 384          // return k + (((base - tmin + 1) * delta) div (delta + skew))
 385          return $k + floor(((self::BOOTSTRAP_BASE - self::BOOTSTRAP_TMIN + 1) * $delta) / ($delta + self::BOOTSTRAP_SKEW));
 386      }
 387  }


Generated : Thu Dec 15 08:20:01 2022 Cross-referenced by PHPXref