RE: Re[2]: [xsl] How can I avoid generating META tag of html?

Subject: RE: Re[2]: [xsl] How can I avoid generating META tag of html?
From: "Julian Reschke" <julian.reschke@xxxxxx>
Date: Thu, 20 Dec 2001 19:43:49 +0100
> From: owner-xsl-list@xxxxxxxxxxxxxxxxxxxxxx
> [mailto:owner-xsl-list@xxxxxxxxxxxxxxxxxxxxxx]On Behalf Of Maxim
> Patramanskij
> Sent: Thursday, December 20, 2001 3:22 PM
> To: Julian Reschke
> Subject: Re[2]: [xsl] How can I avoid generating META tag of html?
>
>
> Hello Julian,
>
> JR> *Why* do you want that?
>
> JR> Hint: don't say it's because you think the information is wrong.
>
> I will try to explain. I have a servlet, which implements site search.
> It generates search result in xml and then tells to transformer to
> apply transformation, having template. As well as I want to have
> output multilanguage, all language-specific information stored in
> Unicode format in .properties files and then selected according to
> sending by browser http header Accept-language.

Sounds good.

> So, depending on browser's language, result's charset could be either
> ISO-8859-1, Cp1251, at the moment. I set encoding value to the

Why would the required *encoding* depend on the *language*? I have a very
similar use case (processing Accept-Language, XML source, XSLT
transformation), but I always send UTF-8 to the client.

> response.setContentType("text/html; charset=" + encoding)
>
> and then do transformation:
>
> transformer.transform(new DOMSource(xmlDoc), new
> StreamResult(response.getWriter()))
>
> If the browser's language was ru (Russian), I see nothing in browser,
> untill manually change browser's encoding from UTF-8 to
> Cyrilic(Windows), which is Cp1251. I do think it happens because
> transformer creates the following in html head:
>
> <META http-equiv="Content-Type" content="text/html; charset=UTF-8">

I think you have a mismatch between the encoding of the response object, and
the encoding used by XSLT. Try setting the right encoding in the top-level
instruction xsl:output.

> To avoid this behavior, I should additionaly do the following:
>
>  transformer.setOutputProperty(OutputKeys.ENCODING, encoding);
>
> before starting the transformation.
>
> But after that, locale specific symbols (Cyrilic) became a sequence of
> Unicode:
>
> <title>&#1056;&#1077;&#1079;&#1091;&#1083;&#1100;&#1090;&#1072;</title>
>
> That's why I'm asking, how to avoid <meta...> generating to let browser
> decide content encoding from http headers.

Well, removing the META tag will not help you at all if the encoding you
specify in the HTTP header isn't the one used by the XSLT engine. Maybe the
encoding you are specifying isn''t supported by the XSLT engine, and it's
falling back to the system encoding?


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread