webmaster@1: <?php
webmaster@1: // $Id: unicode.inc,v 1.29 2007/12/28 12:02:50 dries Exp $
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Indicates an error during check for PHP unicode support.
webmaster@1:  */
webmaster@1: define('UNICODE_ERROR', -1);
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Indicates that standard PHP (emulated) unicode support is being used.
webmaster@1:  */
webmaster@1: define('UNICODE_SINGLEBYTE', 0);
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Indicates that full unicode support with the PHP mbstring extension is being
webmaster@1:  * used.
webmaster@1:  */
webmaster@1: define('UNICODE_MULTIBYTE', 1);
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Wrapper around _unicode_check().
webmaster@1:  */
webmaster@1: function unicode_check() {
webmaster@1:   list($GLOBALS['multibyte']) = _unicode_check();
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Perform checks about Unicode support in PHP, and set the right settings if
webmaster@1:  * needed.
webmaster@1:  *
webmaster@1:  * Because Drupal needs to be able to handle text in various encodings, we do
webmaster@1:  * not support mbstring function overloading. HTTP input/output conversion must
webmaster@1:  * be disabled for similar reasons.
webmaster@1:  *
webmaster@1:  * @param $errors
webmaster@1:  *   Whether to report any fatal errors with form_set_error().
webmaster@1:  */
webmaster@1: function _unicode_check() {
webmaster@1:   // Ensure translations don't break at install time
webmaster@1:   $t = get_t();
webmaster@1: 
webmaster@1:   // Set the standard C locale to ensure consistent, ASCII-only string handling.
webmaster@1:   setlocale(LC_CTYPE, 'C');
webmaster@1: 
webmaster@1:   // Check for outdated PCRE library
webmaster@1:   // Note: we check if U+E2 is in the range U+E0 - U+E1. This test returns TRUE on old PCRE versions.
webmaster@1:   if (preg_match('/[à-á]/u', 'â')) {
webmaster@1:     return array(UNICODE_ERROR, $t('The PCRE library in your PHP installation is outdated. This will cause problems when handling Unicode text. If you are running PHP 4.3.3 or higher, make sure you are using the PCRE library supplied by PHP. Please refer to the <a href="@url">PHP PCRE documentation</a> for more information.', array('@url' => 'http://www.php.net/pcre')));
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Check for mbstring extension
webmaster@1:   if (!function_exists('mb_strlen')) {
webmaster@1:     return array(UNICODE_SINGLEBYTE, $t('Operations on Unicode strings are emulated on a best-effort basis. Install the <a href="@url">PHP mbstring extension</a> for improved Unicode support.', array('@url' => 'http://www.php.net/mbstring')));
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Check mbstring configuration
webmaster@1:   if (ini_get('mbstring.func_overload') != 0) {
webmaster@1:     return array(UNICODE_ERROR, $t('Multibyte string function overloading in PHP is active and must be disabled. Check the php.ini <em>mbstring.func_overload</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));
webmaster@1:   }
webmaster@1:   if (ini_get('mbstring.encoding_translation') != 0) {
webmaster@1:     return array(UNICODE_ERROR, $t('Multibyte string input conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.encoding_translation</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));
webmaster@1:   }
webmaster@1:   if (ini_get('mbstring.http_input') != 'pass') {
webmaster@1:     return array(UNICODE_ERROR, $t('Multibyte string input conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.http_input</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));
webmaster@1:   }
webmaster@1:   if (ini_get('mbstring.http_output') != 'pass') {
webmaster@1:     return array(UNICODE_ERROR, $t('Multibyte string output conversion in PHP is active and must be disabled. Check the php.ini <em>mbstring.http_output</em> setting. Please refer to the <a href="@url">PHP mbstring documentation</a> for more information.', array('@url' => 'http://www.php.net/mbstring')));
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Set appropriate configuration
webmaster@1:   mb_internal_encoding('utf-8');
webmaster@1:   mb_language('uni');
webmaster@1:   return array(UNICODE_MULTIBYTE, '');
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Return Unicode library status and errors.
webmaster@1:  */
webmaster@1: function unicode_requirements() {
webmaster@1:   // Ensure translations don't break at install time
webmaster@1:   $t = get_t();
webmaster@1: 
webmaster@1:   $libraries = array(
webmaster@1:     UNICODE_SINGLEBYTE => $t('Standard PHP'),
webmaster@1:     UNICODE_MULTIBYTE => $t('PHP Mbstring Extension'),
webmaster@1:     UNICODE_ERROR => $t('Error'),
webmaster@1:   );
webmaster@1:   $severities = array(
webmaster@1:     UNICODE_SINGLEBYTE => REQUIREMENT_WARNING,
webmaster@1:     UNICODE_MULTIBYTE => REQUIREMENT_OK,
webmaster@1:     UNICODE_ERROR => REQUIREMENT_ERROR,
webmaster@1:   );
webmaster@1:   list($library, $description) = _unicode_check();
webmaster@1: 
webmaster@1:   $requirements['unicode'] = array(
webmaster@1:     'title' => $t('Unicode library'),
webmaster@1:     'value' => $libraries[$library],
webmaster@1:   );
webmaster@1:   if ($description) {
webmaster@1:     $requirements['unicode']['description'] = $description;
webmaster@1:   }
webmaster@1: 
webmaster@1:   $requirements['unicode']['severity'] = $severities[$library];
webmaster@1: 
webmaster@1:   return $requirements;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Prepare a new XML parser.
webmaster@1:  *
webmaster@1:  * This is a wrapper around xml_parser_create() which extracts the encoding from
webmaster@1:  * the XML data first and sets the output encoding to UTF-8. This function should
webmaster@1:  * be used instead of xml_parser_create(), because PHP 4's XML parser doesn't
webmaster@1:  * check the input encoding itself. "Starting from PHP 5, the input encoding is
webmaster@1:  * automatically detected, so that the encoding parameter specifies only the
webmaster@1:  * output encoding."
webmaster@1:  *
webmaster@1:  * This is also where unsupported encodings will be converted. Callers should
webmaster@1:  * take this into account: $data might have been changed after the call.
webmaster@1:  *
webmaster@1:  * @param &$data
webmaster@1:  *   The XML data which will be parsed later.
webmaster@1:  * @return
webmaster@1:  *   An XML parser object.
webmaster@1:  */
webmaster@1: function drupal_xml_parser_create(&$data) {
webmaster@1:   // Default XML encoding is UTF-8
webmaster@1:   $encoding = 'utf-8';
webmaster@1:   $bom = FALSE;
webmaster@1: 
webmaster@1:   // Check for UTF-8 byte order mark (PHP5's XML parser doesn't handle it).
webmaster@1:   if (!strncmp($data, "\xEF\xBB\xBF", 3)) {
webmaster@1:     $bom = TRUE;
webmaster@1:     $data = substr($data, 3);
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Check for an encoding declaration in the XML prolog if no BOM was found.
webmaster@1:   if (!$bom && ereg('^<\?xml[^>]+encoding="([^"]+)"', $data, $match)) {
webmaster@1:     $encoding = $match[1];
webmaster@1:   }
webmaster@1: 
webmaster@1:   // Unsupported encodings are converted here into UTF-8.
webmaster@1:   $php_supported = array('utf-8', 'iso-8859-1', 'us-ascii');
webmaster@1:   if (!in_array(strtolower($encoding), $php_supported)) {
webmaster@1:     $out = drupal_convert_to_utf8($data, $encoding);
webmaster@1:     if ($out !== FALSE) {
webmaster@1:       $encoding = 'utf-8';
webmaster@1:       $data = ereg_replace('^(<\?xml[^>]+encoding)="([^"]+)"', '\\1="utf-8"', $out);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       watchdog('php', 'Could not convert XML encoding %s to UTF-8.', array('%s' => $encoding), WATCHDOG_WARNING);
webmaster@1:       return 0;
webmaster@1:     }
webmaster@1:   }
webmaster@1: 
webmaster@1:   $xml_parser = xml_parser_create($encoding);
webmaster@1:   xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, 'utf-8');
webmaster@1:   return $xml_parser;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Convert data to UTF-8
webmaster@1:  *
webmaster@1:  * Requires the iconv, GNU recode or mbstring PHP extension.
webmaster@1:  *
webmaster@1:  * @param $data
webmaster@1:  *   The data to be converted.
webmaster@1:  * @param $encoding
webmaster@1:  *   The encoding that the data is in
webmaster@1:  * @return
webmaster@1:  *   Converted data or FALSE.
webmaster@1:  */
webmaster@1: function drupal_convert_to_utf8($data, $encoding) {
webmaster@1:   if (function_exists('iconv')) {
webmaster@1:     $out = @iconv($encoding, 'utf-8', $data);
webmaster@1:   }
webmaster@1:   else if (function_exists('mb_convert_encoding')) {
webmaster@1:     $out = @mb_convert_encoding($data, 'utf-8', $encoding);
webmaster@1:   }
webmaster@1:   else if (function_exists('recode_string')) {
webmaster@1:     $out = @recode_string($encoding .'..utf-8', $data);
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     watchdog('php', 'Unsupported encoding %s. Please install iconv, GNU recode or mbstring for PHP.', array('%s' => $encoding), WATCHDOG_ERROR);
webmaster@1:     return FALSE;
webmaster@1:   }
webmaster@1: 
webmaster@1:   return $out;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Truncate a UTF-8-encoded string safely to a number of bytes.
webmaster@1:  *
webmaster@1:  * If the end position is in the middle of a UTF-8 sequence, it scans backwards
webmaster@1:  * until the beginning of the byte sequence.
webmaster@1:  *
webmaster@1:  * Use this function whenever you want to chop off a string at an unsure
webmaster@1:  * location. On the other hand, if you're sure that you're splitting on a
webmaster@1:  * character boundary (e.g. after using strpos() or similar), you can safely use
webmaster@1:  * substr() instead.
webmaster@1:  *
webmaster@1:  * @param $string
webmaster@1:  *   The string to truncate.
webmaster@1:  * @param $len
webmaster@1:  *   An upper limit on the returned string length.
webmaster@1:  * @return
webmaster@1:  *   The truncated string.
webmaster@1:  */
webmaster@1: function drupal_truncate_bytes($string, $len) {
webmaster@1:   if (strlen($string) <= $len) {
webmaster@1:     return $string;
webmaster@1:   }
webmaster@1:   if ((ord($string[$len]) < 0x80) || (ord($string[$len]) >= 0xC0)) {
webmaster@1:     return substr($string, 0, $len);
webmaster@1:   }
webmaster@1:   while (--$len >= 0 && ord($string[$len]) >= 0x80 && ord($string[$len]) < 0xC0) {};
webmaster@1:   return substr($string, 0, $len);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Truncate a UTF-8-encoded string safely to a number of characters.
webmaster@1:  *
webmaster@1:  * @param $string
webmaster@1:  *   The string to truncate.
webmaster@1:  * @param $len
webmaster@1:  *   An upper limit on the returned string length.
webmaster@1:  * @param $wordsafe
webmaster@1:  *   Flag to truncate at last space within the upper limit. Defaults to FALSE.
webmaster@1:  * @param $dots
webmaster@1:  *   Flag to add trailing dots. Defaults to FALSE.
webmaster@1:  * @return
webmaster@1:  *   The truncated string.
webmaster@1:  */
webmaster@1: function truncate_utf8($string, $len, $wordsafe = FALSE, $dots = FALSE) {
webmaster@1: 
webmaster@1:   if (drupal_strlen($string) <= $len) {
webmaster@1:     return $string;
webmaster@1:   }
webmaster@1: 
webmaster@1:   if ($dots) {
webmaster@1:     $len -= 4;
webmaster@1:   }
webmaster@1: 
webmaster@1:   if ($wordsafe) {
webmaster@1:     $string = drupal_substr($string, 0, $len + 1); // leave one more character
webmaster@1:     if ($last_space = strrpos($string, ' ')) { // space exists AND is not on position 0
webmaster@1:       $string = substr($string, 0, $last_space);
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       $string = drupal_substr($string, 0, $len);
webmaster@1:     }
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     $string = drupal_substr($string, 0, $len);
webmaster@1:   }
webmaster@1: 
webmaster@1:   if ($dots) {
webmaster@1:     $string .= ' ...';
webmaster@1:   }
webmaster@1: 
webmaster@1:   return $string;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Encodes MIME/HTTP header values that contain non-ASCII, UTF-8 encoded
webmaster@1:  * characters.
webmaster@1:  *
webmaster@1:  * For example, mime_header_encode('tést.txt') returns "=?UTF-8?B?dMOpc3QudHh0?=".
webmaster@1:  *
webmaster@1:  * See http://www.rfc-editor.org/rfc/rfc2047.txt for more information.
webmaster@1:  *
webmaster@1:  * Notes:
webmaster@1:  * - Only encode strings that contain non-ASCII characters.
webmaster@1:  * - We progressively cut-off a chunk with truncate_utf8(). This is to ensure
webmaster@1:  *   each chunk starts and ends on a character boundary.
webmaster@1:  * - Using \n as the chunk separator may cause problems on some systems and may
webmaster@1:  *   have to be changed to \r\n or \r.
webmaster@1:  */
webmaster@1: function mime_header_encode($string) {
webmaster@1:   if (preg_match('/[^\x20-\x7E]/', $string)) {
webmaster@1:     $chunk_size = 47; // floor((75 - strlen("=?UTF-8?B??=")) * 0.75);
webmaster@1:     $len = strlen($string);
webmaster@1:     $output = '';
webmaster@1:     while ($len > 0) {
webmaster@1:       $chunk = drupal_truncate_bytes($string, $chunk_size);
webmaster@1:       $output .= ' =?UTF-8?B?'. base64_encode($chunk) ."?=\n";
webmaster@1:       $c = strlen($chunk);
webmaster@1:       $string = substr($string, $c);
webmaster@1:       $len -= $c;
webmaster@1:     }
webmaster@1:     return trim($output);
webmaster@1:   }
webmaster@1:   return $string;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Complement to mime_header_encode
webmaster@1:  */
webmaster@1: function mime_header_decode($header) {
webmaster@1:   // First step: encoded chunks followed by other encoded chunks (need to collapse whitespace)
webmaster@1:   $header = preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=\s+(?==\?)/', '_mime_header_decode', $header);
webmaster@1:   // Second step: remaining chunks (do not collapse whitespace)
webmaster@1:   return preg_replace_callback('/=\?([^?]+)\?(Q|B)\?([^?]+|\?(?!=))\?=/', '_mime_header_decode', $header);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function to mime_header_decode
webmaster@1:  */
webmaster@1: function _mime_header_decode($matches) {
webmaster@1:   // Regexp groups:
webmaster@1:   // 1: Character set name
webmaster@1:   // 2: Escaping method (Q or B)
webmaster@1:   // 3: Encoded data
webmaster@1:   $data = ($matches[2] == 'B') ? base64_decode($matches[3]) : str_replace('_', ' ', quoted_printable_decode($matches[3]));
webmaster@1:   if (strtolower($matches[1]) != 'utf-8') {
webmaster@1:     $data = drupal_convert_to_utf8($data, $matches[1]);
webmaster@1:   }
webmaster@1:   return $data;
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Decode all HTML entities (including numerical ones) to regular UTF-8 bytes.
webmaster@1:  * Double-escaped entities will only be decoded once ("&amp;lt;" becomes "&lt;", not "<").
webmaster@1:  *
webmaster@1:  * @param $text
webmaster@1:  *   The text to decode entities in.
webmaster@1:  * @param $exclude
webmaster@1:  *   An array of characters which should not be decoded. For example,
webmaster@1:  *   array('<', '&', '"'). This affects both named and numerical entities.
webmaster@1:  */
webmaster@1: function decode_entities($text, $exclude = array()) {
webmaster@1:   static $table;
webmaster@1:   // We store named entities in a table for quick processing.
webmaster@1:   if (!isset($table)) {
webmaster@1:     // Get all named HTML entities.
webmaster@1:     $table = array_flip(get_html_translation_table(HTML_ENTITIES));
webmaster@1:     // PHP gives us ISO-8859-1 data, we need UTF-8.
webmaster@1:     $table = array_map('utf8_encode', $table);
webmaster@1:     // Add apostrophe (XML)
webmaster@1:     $table['&apos;'] = "'";
webmaster@1:   }
webmaster@1:   $newtable = array_diff($table, $exclude);
webmaster@1: 
webmaster@1:   // Use a regexp to select all entities in one pass, to avoid decoding double-escaped entities twice.
webmaster@1:   return preg_replace('/&(#x?)?([A-Za-z0-9]+);/e', '_decode_entities("$1", "$2", "$0", $newtable, $exclude)', $text);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for decode_entities
webmaster@1:  */
webmaster@1: function _decode_entities($prefix, $codepoint, $original, &$table, &$exclude) {
webmaster@1:   // Named entity
webmaster@1:   if (!$prefix) {
webmaster@1:     if (isset($table[$original])) {
webmaster@1:       return $table[$original];
webmaster@1:     }
webmaster@1:     else {
webmaster@1:       return $original;
webmaster@1:     }
webmaster@1:   }
webmaster@1:   // Hexadecimal numerical entity
webmaster@1:   if ($prefix == '#x') {
webmaster@1:     $codepoint = base_convert($codepoint, 16, 10);
webmaster@1:   }
webmaster@1:   // Decimal numerical entity (strip leading zeros to avoid PHP octal notation)
webmaster@1:   else {
webmaster@1:     $codepoint = preg_replace('/^0+/', '', $codepoint);
webmaster@1:   }
webmaster@1:   // Encode codepoint as UTF-8 bytes
webmaster@1:   if ($codepoint < 0x80) {
webmaster@1:     $str = chr($codepoint);
webmaster@1:   }
webmaster@1:   else if ($codepoint < 0x800) {
webmaster@1:     $str = chr(0xC0 | ($codepoint >> 6))
webmaster@1:          . chr(0x80 | ($codepoint & 0x3F));
webmaster@1:   }
webmaster@1:   else if ($codepoint < 0x10000) {
webmaster@1:     $str = chr(0xE0 | ( $codepoint >> 12))
webmaster@1:          . chr(0x80 | (($codepoint >> 6) & 0x3F))
webmaster@1:          . chr(0x80 | ( $codepoint       & 0x3F));
webmaster@1:   }
webmaster@1:   else if ($codepoint < 0x200000) {
webmaster@1:     $str = chr(0xF0 | ( $codepoint >> 18))
webmaster@1:          . chr(0x80 | (($codepoint >> 12) & 0x3F))
webmaster@1:          . chr(0x80 | (($codepoint >> 6)  & 0x3F))
webmaster@1:          . chr(0x80 | ( $codepoint        & 0x3F));
webmaster@1:   }
webmaster@1:   // Check for excluded characters
webmaster@1:   if (in_array($str, $exclude)) {
webmaster@1:     return $original;
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     return $str;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Count the amount of characters in a UTF-8 string. This is less than or
webmaster@1:  * equal to the byte count.
webmaster@1:  */
webmaster@1: function drupal_strlen($text) {
webmaster@1:   global $multibyte;
webmaster@1:   if ($multibyte == UNICODE_MULTIBYTE) {
webmaster@1:     return mb_strlen($text);
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     // Do not count UTF-8 continuation bytes.
webmaster@1:     return strlen(preg_replace("/[\x80-\xBF]/", '', $text));
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Uppercase a UTF-8 string.
webmaster@1:  */
webmaster@1: function drupal_strtoupper($text) {
webmaster@1:   global $multibyte;
webmaster@1:   if ($multibyte == UNICODE_MULTIBYTE) {
webmaster@1:     return mb_strtoupper($text);
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     // Use C-locale for ASCII-only uppercase
webmaster@1:     $text = strtoupper($text);
webmaster@1:     // Case flip Latin-1 accented letters
webmaster@1:     $text = preg_replace_callback('/\xC3[\xA0-\xB6\xB8-\xBE]/', '_unicode_caseflip', $text);
webmaster@1:     return $text;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Lowercase a UTF-8 string.
webmaster@1:  */
webmaster@1: function drupal_strtolower($text) {
webmaster@1:   global $multibyte;
webmaster@1:   if ($multibyte == UNICODE_MULTIBYTE) {
webmaster@1:     return mb_strtolower($text);
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     // Use C-locale for ASCII-only lowercase
webmaster@1:     $text = strtolower($text);
webmaster@1:     // Case flip Latin-1 accented letters
webmaster@1:     $text = preg_replace_callback('/\xC3[\x80-\x96\x98-\x9E]/', '_unicode_caseflip', $text);
webmaster@1:     return $text;
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Helper function for case conversion of Latin-1.
webmaster@1:  * Used for flipping U+C0-U+DE to U+E0-U+FD and back.
webmaster@1:  */
webmaster@1: function _unicode_caseflip($matches) {
webmaster@1:   return $matches[0][0] . chr(ord($matches[0][1]) ^ 32);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Capitalize the first letter of a UTF-8 string.
webmaster@1:  */
webmaster@1: function drupal_ucfirst($text) {
webmaster@1:   // Note: no mbstring equivalent!
webmaster@1:   return drupal_strtoupper(drupal_substr($text, 0, 1)) . drupal_substr($text, 1);
webmaster@1: }
webmaster@1: 
webmaster@1: /**
webmaster@1:  * Cut off a piece of a string based on character indices and counts. Follows
webmaster@1:  * the same behavior as PHP's own substr() function.
webmaster@1:  *
webmaster@1:  * Note that for cutting off a string at a known character/substring
webmaster@1:  * location, the usage of PHP's normal strpos/substr is safe and
webmaster@1:  * much faster.
webmaster@1:  */
webmaster@1: function drupal_substr($text, $start, $length = NULL) {
webmaster@1:   global $multibyte;
webmaster@1:   if ($multibyte == UNICODE_MULTIBYTE) {
webmaster@1:     return $length === NULL ? mb_substr($text, $start) : mb_substr($text, $start, $length);
webmaster@1:   }
webmaster@1:   else {
webmaster@1:     $strlen = strlen($text);
webmaster@1:     // Find the starting byte offset
webmaster@1:     $bytes = 0;
webmaster@1:     if ($start > 0) {
webmaster@1:       // Count all the continuation bytes from the start until we have found
webmaster@1:       // $start characters
webmaster@1:       $bytes = -1; $chars = -1;
webmaster@1:       while ($bytes < $strlen && $chars < $start) {
webmaster@1:         $bytes++;
webmaster@1:         $c = ord($text[$bytes]);
webmaster@1:         if ($c < 0x80 || $c >= 0xC0) {
webmaster@1:           $chars++;
webmaster@1:         }
webmaster@1:       }
webmaster@1:     }
webmaster@1:     else if ($start < 0) {
webmaster@1:       // Count all the continuation bytes from the end until we have found
webmaster@1:       // abs($start) characters
webmaster@1:       $start = abs($start);
webmaster@1:       $bytes = $strlen; $chars = 0;
webmaster@1:       while ($bytes > 0 && $chars < $start) {
webmaster@1:         $bytes--;
webmaster@1:         $c = ord($text[$bytes]);
webmaster@1:         if ($c < 0x80 || $c >= 0xC0) {
webmaster@1:           $chars++;
webmaster@1:         }
webmaster@1:       }
webmaster@1:     }
webmaster@1:     $istart = $bytes;
webmaster@1: 
webmaster@1:     // Find the ending byte offset
webmaster@1:     if ($length === NULL) {
webmaster@1:       $bytes = $strlen - 1;
webmaster@1:     }
webmaster@1:     else if ($length > 0) {
webmaster@1:       // Count all the continuation bytes from the starting index until we have
webmaster@1:       // found $length + 1 characters. Then backtrack one byte.
webmaster@1:       $bytes = $istart; $chars = 0;
webmaster@1:       while ($bytes < $strlen && $chars < $length) {
webmaster@1:         $bytes++;
webmaster@1:         $c = ord($text[$bytes]);
webmaster@1:         if ($c < 0x80 || $c >= 0xC0) {
webmaster@1:           $chars++;
webmaster@1:         }
webmaster@1:       }
webmaster@1:       $bytes--;
webmaster@1:     }
webmaster@1:     else if ($length < 0) {
webmaster@1:       // Count all the continuation bytes from the end until we have found
webmaster@1:       // abs($length) characters
webmaster@1:       $length = abs($length);
webmaster@1:       $bytes = $strlen - 1; $chars = 0;
webmaster@1:       while ($bytes >= 0 && $chars < $length) {
webmaster@1:         $c = ord($text[$bytes]);
webmaster@1:         if ($c < 0x80 || $c >= 0xC0) {
webmaster@1:           $chars++;
webmaster@1:         }
webmaster@1:         $bytes--;
webmaster@1:       }
webmaster@1:     }
webmaster@1:     $iend = $bytes;
webmaster@1: 
webmaster@1:     return substr($text, $istart, max(0, $iend - $istart + 1));
webmaster@1:   }
webmaster@1: }
webmaster@1: 
webmaster@1: