Description:
------------
Per summary, backreferences are not properly replaced if the regex encoding is any of the UTF-32 flavors (BE, LE, UCS-4).  All 3 parameters (pattern, replace, subject) are being presented in the correct UTF-32 flavor; replace is being returned unmodified.

Test script:
---------------
$regexEncoding = 'UTF-32';
mb_regex_encoding( $regexEncoding );
var_export( mb_ereg_replace( mb_convert_encoding( '.+', $regexEncoding ), mb_convert_encoding( "the \\0", $regexEncoding ), mb_convert_encoding( 'dog', $regexEncoding ) ) );

Expected result:
----------------
'' . "\0" . '' . "\0" . '' . "\0" . 't' . "\0" . '' . "\0" . '' . "\0" . 'h' . "\0" . '' . "\0" . '' . "\0" . 'e' . "\0" . '' . "\0" . '' . "\0" . ' ' . "\0" . '' . "\0" . '' . "\0" . 'd' . "\0" . '' . "\0" . '' . "\0" . 'o' . "\0" . '' . "\0" . '' . "\0" . 'g'

e.g., "the dog" in UTF-32 encoding

Actual result:
--------------
'' . "\0" . '' . "\0" . '' . "\0" . 't' . "\0" . '' . "\0" . '' . "\0" . 'h' . "\0" . '' . "\0" . '' . "\0" . 'e' . "\0" . '' . "\0" . '' . "\0" . ' ' . "\0" . '' . "\0" . '' . "\0" . '\\' . "\0" . '' . "\0" . '' . "\0" . '0'

e.g., "\\0" in UTF-32 encoding