Has someone got around the problem of "undecoded UTF-8 will give garbage when decoded". I know it is generated within HTML::Parser and can be fixed executing utf8_mode on the HTML::Parser object. But how can I access the HTML::Parser object from WWW::Mechanize ?