Cocoon is heavily internationalized but we fail to do one thing: signal the proper encoding to the user-agent thru HTTP headers, which is the most reliable way of doing it.
the current *hack* is to use <meta> tags in the HTML stream,
Ew!
these are interpreted by the HTTP server stack and transfered as HTTP headers. but this creates many problems and concern mixes.
Vadim suggested to set the headers from the serializers, but I think there is a better alternative.
So I propose to add the method
getEncoding()
to the interface
org.apache.cocoon.sitemap
Why sitemap would ever know anything about encoding?
There are two parts to the encoding problem: decoding incoming request and encoding outgoing response. Request encoding can be set by SetCharacterEncodingAction or by anything else via request.setCharacterEncoding() method. Or, every request parameter can be decoded independently. Response encoding directly depends on the encoding parameter set to the serializer from the sitemap.
And, any of these are totally independent from the internationalization. Internationalization affects language used to produce output, but not how the text in this language is encoded (UTF8, UTF16, ISO-1859-1, what-have-you).
So, if you to put encoding into sitemap... You will have to disable serializer configuration and request configuration and force sitemap encoding onto request / response. Is this what you are proposing? If yes... IMHO, it makes more sence to have this parameter of the pipeline but not whole sitemap. But I am not convinced that it's sitemap's responsibility to worry about encoding (from SoC POV).
Vadim
