Expose the encoding of a document

[SimonSapin]

Based on various information provided by the API’s user (see #120) and in the input stream (BOM, <meta charset>), the parser decides which character encoding to use.

The used encoding for a parsed document should be exposed somehow. It is needed eg. as a fallback encoding for scripts and stylesheets referred to by the document.

[SimonSapin]

This is actually accessible as HTML5Parser.tokenizer.stream.charEncoding.