XHTML parser ignores attributes of wikimodel-freestanding images

Currently wikimodel-freestanding CSS class has priority over image attributes. As a consequence image attributes are ignored when this class is present. IMO the XHTML parser should check the presence of wikimodel-freestanding class only if the image doesn't have attributes.

wikimodel-freestanding CSS class is used to distinguish between image:reference and [[image:reference]] (without parameters) so it's useless when the image has parameters (you have to use the [[..]] notation).