On 9 January 2012 11:36, Thorsten Scherler <scher...@gmail.com> wrote: > On Mon, 2012-01-09 at 08:32 +0100, Robby Pelssers wrote: >> Hi Thorsten, >> >> Adding <meta> in general is not a concern faik but setting the correct >> encoding is. >> >> Examples are >> <?xml version="1.0" encoding="UTF-8"?> for xml files > > That is correct for the doc declaration. > >> And >> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"/> for >> html files > > nupp, that tag may be needed to be valid html5 but that is not the > concern of the serializer but the prior transformation process. > >> >> So I was only referring to setting the correct encoding which can be >> configured as a Serializer property. > > Yes but that only goes in the PI and is used for the serialization.
Not really convinced, chiefly for reasons of separation of concerns. Given that throughout the pipeline the XML is being held in java's unicode strings, IMO the only component that should need to worry about the charset being used to serialise the output should be the serialiser that's doing it, otherwise you can end up with a document using one charset that claims inside to be a different one. If you're happy to leave it to the serialiser to insert the PI in the output (including the charset) rather than having it already in the pipeline's XML stream (e.g. inserted by xsl:processing-instruction in an XSLT template), and happy to let the the HTML serialiser insert the doctype rather than having it already in the pipeline's stream, then why shouldn't the HTML/XHTML serialiser also insert the meta tag specifying the charset? In an ideal world, we wouldn't even have to specify a particular encoding on the serialiser either - there'd be a default configured somewhere, but it would select an appropriate one dynamically at the time of output based on the Accept-Charset request header sent by the browser... and why should the earlier part of the pipeline also need to worry about that? Andy. > > salu2 > >> >> Robby >> >> >> -----Original Message----- >> From: Thorsten Scherler [mailto:scher...@gmail.com] >> Sent: Sunday, January 08, 2012 10:28 PM >> To: dev@cocoon.apache.org >> Subject: RE: HTML5 serializer >> >> On Fri, 2012-01-06 at 19:56 +0100, Robby Pelssers wrote: >> > .... >> >> > So we’re almost there. Do you have any suggestion how to accomplish >> > using the correct <meta charset=”utf-8”/> ?? Or do you think that’s >> > not worth the effort? >> >> Hmm, actually that is not the concern of the serializer at all. The >> serializer merely adds DOCTYPE PI and not much more. So <meta> is >> nothing the serializer should add. >> >> salu2 >> > > -- > Thorsten Scherler <thorsten.at.apache.org> > codeBusters S.L. - web based systems > <consulting, training and solutions> > http://www.codebusters.es/