comparison vendor/symfony/polyfill-mbstring/Mbstring.php @ 12:7a779792577d

Update Drupal core to v8.4.5 (via Composer)

  * - mb_chr                  - Returns a specific character from its Unicode code point
  * - mb_convert_encoding     - Convert character encoding
  * - mb_convert_variables    - Convert character code in variable(s)
  * - mb_decode_mimeheader    - Decode string in MIME header field
  * - mb_encode_mimeheader    - Encode string for MIME header XXX NATIVE IMPLEMENTATION IS REALLY BUGGED
+ * - mb_decode_numericentity - Decode HTML numeric string reference to character
+ * - mb_encode_numericentity - Encode character to HTML numeric string reference
  * - mb_convert_case         - Perform case folding on a string
+ * - mb_detect_encoding      - Detect character encoding
  * - mb_get_info             - Get internal settings of mbstring
  * - mb_http_input           - Detect HTTP input character encoding
  * - mb_http_output          - Set/Get HTTP output character encoding
  * - mb_internal_encoding    - Set/Get internal character encoding
  * - mb_list_encodings       - Returns an array of all supported encodings

  * - mb_strwidth             - Return width of string
  * - mb_substr_count         - Count the number of substring occurrences
  *
  * Not implemented:
  * - mb_convert_kana         - Convert "kana" one from another ("zen-kaku", "han-kaku" and more)
- * - mb_decode_numericentity - Decode HTML numeric string reference to character
- * - mb_encode_numericentity - Encode character to HTML numeric string reference
  * - mb_ereg_*               - Regular expression with multibyte support
  * - mb_parse_str            - Parse GET/POST/COOKIE data and set global variable
  * - mb_preferred_mime_name  - Get MIME charset string
  * - mb_regex_encoding       - Returns current encoding for multibyte regex as string
  * - mb_regex_set_options    - Set/Get the default options for mbregex functions

     public static function mb_encode_mimeheader($s, $charset = null, $transferEncoding = null, $linefeed = null, $indent = null)
     {
         trigger_error('mb_encode_mimeheader() is bugged. Please use iconv_mime_encode() instead', E_USER_WARNING);
     }
 
-    public static function mb_convert_case($s, $mode, $encoding = null)
-    {
-        if ('' === $s .= '') {
+    public static function mb_decode_numericentity($s, $convmap, $encoding = null)
+    {
+        if (null !== $s && !is_scalar($s) && !(is_object($s) && method_exists($s, '__toString'))) {
+            trigger_error('mb_decode_numericentity() expects parameter 1 to be string, '.gettype($s).' given', E_USER_WARNING);
+            return null;
+        }
+
+        if (!is_array($convmap) || !$convmap) {
+            return false;
+        }
+
+        if (null !== $encoding && !is_scalar($encoding)) {
+            trigger_error('mb_decode_numericentity() expects parameter 3 to be string, '.gettype($s).' given', E_USER_WARNING);
+            return '';  // Instead of null (cf. mb_encode_numericentity).
+        }
+
+        $s = (string) $s;
+        if ('' === $s) {
             return '';
         }
 
         $encoding = self::getEncoding($encoding);
 
         if ('UTF-8' === $encoding) {
             $encoding = null;
+            if (!preg_match('//u', $s)) {
+                $s = @iconv('UTF-8', 'UTF-8//IGNORE', $s);
+            }
+        } else {
+            $s = iconv($encoding, 'UTF-8//IGNORE', $s);
+        }
+
+        $cnt = floor(count($convmap) / 4) * 4;
+
+        for ($i = 0; $i < $cnt; $i += 4) {
+            // collector_decode_htmlnumericentity ignores $convmap[$i + 3]
+            $convmap[$i] += $convmap[$i + 2];
+            $convmap[$i + 1] += $convmap[$i + 2];
+        }
+
+        $s = preg_replace_callback('/&#(?:0*([0-9]+)|x0*([0-9a-fA-F]+))(?!&);?/', function (array $m) use ($cnt, $convmap) {
+            $c = isset($m[2]) ? (int) hexdec($m[2]) : $m[1];
+            for ($i = 0; $i < $cnt; $i += 4) {
+                if ($c >= $convmap[$i] && $c <= $convmap[$i + 1]) {
+                    return Mbstring::mb_chr($c - $convmap[$i + 2]);
+                }
+            }
+            return $m[0];
+        }, $s);
+
+        if (null === $encoding) {
+            return $s;
+        }
+
+        return iconv('UTF-8', $encoding.'//IGNORE', $s);
+    }
+
+    public static function mb_encode_numericentity($s, $convmap, $encoding = null, $is_hex = false)
+    {
+        if (null !== $s && !is_scalar($s) && !(is_object($s) && method_exists($s, '__toString'))) {
+            trigger_error('mb_encode_numericentity() expects parameter 1 to be string, '.gettype($s).' given', E_USER_WARNING);
+            return null;
+        }
+
+        if (!is_array($convmap) || !$convmap) {
+            return false;
+        }
+
+        if (null !== $encoding && !is_scalar($encoding)) {
+            trigger_error('mb_encode_numericentity() expects parameter 3 to be string, '.gettype($s).' given', E_USER_WARNING);
+            return null;  // Instead of '' (cf. mb_decode_numericentity).
+        }
+
+        if (null !== $is_hex && !is_scalar($is_hex)) {
+            trigger_error('mb_encode_numericentity() expects parameter 4 to be boolean, '.gettype($s).' given', E_USER_WARNING);
+            return null;
+        }
+
+        $s = (string) $s;
+        if ('' === $s) {
+            return '';
+        }
+
+        $encoding = self::getEncoding($encoding);
+
+        if ('UTF-8' === $encoding) {
+            $encoding = null;
+            if (!preg_match('//u', $s)) {
+                $s = @iconv('UTF-8', 'UTF-8//IGNORE', $s);
+            }
+        } else {
+            $s = iconv($encoding, 'UTF-8//IGNORE', $s);
+        }
+
+        static $ulenMask = array("\xC0" => 2, "\xD0" => 2, "\xE0" => 3, "\xF0" => 4);
+
+        $cnt = floor(count($convmap) / 4) * 4;
+        $i = 0;
+        $len = strlen($s);
+        $result = '';
+
+        while ($i < $len) {
+            $ulen = $s[$i] < "\x80" ? 1 : $ulenMask[$s[$i] & "\xF0"];
+            $uchr = substr($s, $i, $ulen);
+            $i += $ulen;
+            $c = self::mb_ord($uchr);
+
+            for ($j = 0; $j < $cnt; $j += 4) {
+                if ($c >= $convmap[$j] && $c <= $convmap[$j + 1]) {
+                    $cOffset = ($c + $convmap[$j + 2]) & $convmap[$j + 3];
+                    $result .= $is_hex ? sprintf('&#x%X;', $cOffset) : '&#'.$cOffset.';';
+                    continue 2;
+                }
+            }
+            $result .= $uchr;
+        }
+
+        if (null === $encoding) {
+            return $result;
+        }
+
+        return iconv('UTF-8', $encoding.'//IGNORE', $result);
+    }
+
+    public static function mb_convert_case($s, $mode, $encoding = null)
+    {
+        $s = (string) $s;
+        if ('' === $s) {
+            return '';
+        }
+
+        $encoding = self::getEncoding($encoding);
+
+        if ('UTF-8' === $encoding) {
+            $encoding = null;
+            if (!preg_match('//u', $s)) {
+                $s = @iconv('UTF-8', 'UTF-8//IGNORE', $s);
+            }
         } else {
             $s = iconv($encoding, 'UTF-8//IGNORE', $s);
         }
 
         if (MB_CASE_TITLE == $mode) {

         return true;
     }
 
     public static function mb_strlen($s, $encoding = null)
     {
-        switch ($encoding = self::getEncoding($encoding)) {
-            case 'ASCII':
-            case 'CP850':
-                return strlen($s);
+        $encoding = self::getEncoding($encoding);
+        if ('CP850' === $encoding || 'ASCII' === $encoding) {
+            return strlen($s);
         }
 
         return @iconv_strlen($s, $encoding);
     }
 
     public static function mb_strpos($haystack, $needle, $offset = 0, $encoding = null)
     {
         $encoding = self::getEncoding($encoding);
-
-        if ('' === $needle .= '') {
+        if ('CP850' === $encoding || 'ASCII' === $encoding) {
+            return strpos($haystack, $needle, $offset);
+        }
+
+        $needle = (string) $needle;
+        if ('' === $needle) {
             trigger_error(__METHOD__.': Empty delimiter', E_USER_WARNING);
 
             return false;
         }
 

     }
 
     public static function mb_strrpos($haystack, $needle, $offset = 0, $encoding = null)
     {
         $encoding = self::getEncoding($encoding);
+        if ('CP850' === $encoding || 'ASCII' === $encoding) {
+            return strrpos($haystack, $needle, $offset);
+        }
 
         if ($offset != (int) $offset) {
             $offset = 0;
         } elseif ($offset = (int) $offset) {
             if ($offset < 0) {

     }
 
     public static function mb_substr($s, $start, $length = null, $encoding = null)
     {
         $encoding = self::getEncoding($encoding);
+        if ('CP850' === $encoding || 'ASCII' === $encoding) {
+            return substr($s, $start, null === $length ? 2147483647 : $length);
+        }
 
         if ($start < 0) {
             $start = iconv_strlen($s, $encoding) + $start;
             if ($start < 0) {
                 $start = 0;

             if ($length < 0) {
                 return '';
             }
         }
 
-        return iconv_substr($s, $start, $length, $encoding).'';
+        return (string) iconv_substr($s, $start, $length, $encoding);
     }
 
     public static function mb_stripos($haystack, $needle, $offset = 0, $encoding = null)
     {
         $haystack = self::mb_convert_case($haystack, self::MB_CASE_FOLD, $encoding);

     }
 
     public static function mb_strrchr($haystack, $needle, $part = false, $encoding = null)
     {
         $encoding = self::getEncoding($encoding);
+        if ('CP850' === $encoding || 'ASCII' === $encoding) {
+            return strrchr($haystack, $needle, $part);
+        }
         $needle = self::mb_substr($needle, 0, 1, $encoding);
         $pos = iconv_strrpos($haystack, $needle, $encoding);
 
         return self::getSubpart($pos, $part, $haystack, $encoding);
     }

         }
 
         return self::mb_substr($haystack, $pos, null, $encoding);
     }
 
-    private static function html_encoding_callback($m)
+    private static function html_encoding_callback(array $m)
     {
         $i = 1;
         $entities = '';
         $m = unpack('C*', htmlentities($m[0], ENT_COMPAT, 'UTF-8'));
 

         }
 
         return $entities;
     }
 
-    private static function title_case_lower($s)
+    private static function title_case_lower(array $s)
     {
         return self::mb_convert_case($s[0], MB_CASE_LOWER, 'UTF-8');
     }
 
-    private static function title_case_upper($s)
+    private static function title_case_upper(array $s)
     {
         return self::mb_convert_case($s[0], MB_CASE_UPPER, 'UTF-8');
     }
 
     private static function getData($file)