Description:
------------
It may be a misunderstanding of mine, but xml_parse doesn't seem to resolve 'simple' entities like < > or &. According to the doc, and one other bug report, I think it should.
While other parsers (XMLReader, SimpleXML, DOM, ...) do.

In numerical form, it does work (>).

I tested with libxml2 versions 2.7.1 and 2.7.2, and expat 2.0.1.

I found this when doing SOAP calls, I couldn't pass the value "<" to a PHP SOAP server (which is using NuSOAP).

I don't have a workaround for the moment.

Reproduce code:
---------------
function character_data($parser, $data){
        print("character_data: $data\n");
}

function parsestring( $strg ) {
        $parser = xml_parser_create("UTF-8");
        xml_set_character_data_handler($parser,'character_data');
        xml_parse($parser,$strg,true);
        xml_parser_free($parser);

        $reader = new XMLReader();
        $reader->XML( $strg );
        while( $reader->read() ) {
                if( $reader->value ) print( $reader->name . ": " . $reader->value . "\n");
        }
}

$xmlcdata = "<element><![CDATA[A>B&C<D]]></element>";
$xmlescaped = "<element>A&gt;B&amp;C&lt;D</element>";
$xmlnumeric = "<element>A&#x3e;B&#x26;C&#x3c;D</element>";

print("parsing $xmlcdata \n");
parsestring( $xmlcdata );
print("\nparsing $xmlescaped \n");
parsestring( $xmlescaped );
print("\nparsing $xmlnumeric \n");
parsestring( $xmlnumeric );


Expected result:
----------------
parsing <element><![CDATA[A>B&C<D]]></element>
character_data: A>B&C<D
#cdata-section: A>B&C<D

parsing <element>A&gt;B&amp;C&lt;D</element>
character_data: A
character_data: >
character_data: B
character_data: &
character_data: C
character_data: <
character_data: D
#text: A>B&C<D

parsing <element>A&#x3e;B&#x26;C&#x3c;D</element>
character_data: A
character_data: >
character_data: B
character_data: &
character_data: C
character_data: <
character_data: D
#text: A>B&C<D


Actual result:
--------------
parsing <element><![CDATA[A>B&C<D]]></element>
character_data: A>B&C<D
#cdata-section: A>B&C<D

parsing <element>A&gt;B&amp;C&lt;D</element>
character_data: A
character_data: B
character_data: C
character_data: D
#text: A>B&C<D

parsing <element>A&#x3e;B&#x26;C&#x3c;D</element>
character_data: A
character_data: >
character_data: B
character_data: &
character_data: C
character_data: <
character_data: D
#text: A>B&C<D