Named character references in HTML5 kill parsing

[kaefer3000]

Hi,

parsing http://schema.org/Person with rapper, I stumbled upon the following problem and I guess it is caused by librdfa: The page containes named character references (see HTML5 spec) such as   and because HTML5 does not come with a DTD, the ones exceeding the default XML escapes are considered undefined thus stopping the parsing with an error:
rapper: Error - - XML parser error: Entity 'nbsp' not defined
Yet, those escapes are allowed in HTML5 and around and it would therefore be great if librdfa could support them.

Cheers,

Tobias

[dajobe]

There's not really work going on for librdfa. I mostly fix minor bugs and portability stuff for raptor.

[kaefer3000]

Ah ok, but maybe it is not so much of a big deal to fix it. For example, there is a JSON file with all the named character references, maybe this information can be supplied to the xml parser, or the parser can be told to ignore that kind of errors. Also, even changing the doctype to HTML4 brings the parser from 11 to 64 triples (of 239 achieved with pyrdfa).