Re: [xsl] using xsl:output-character to render characters in 2 ways

Subject: Re: [xsl] using xsl:output-character to render characters in 2 ways
From: Tom T <tomw289@xxxxxxxxxxxxxx>
Date: Thu, 24 Dec 2009 10:21:30 +0000
So, if I am forced to generate XHTML, but also forced to serve it as
HTML then I am faced with this problem. This seems fair enough.

However, if I am generating XHTML, serving it as XHTML but wish to
begin a CDATA section half way through a passage of text then I am
doing everything correctly but also face a problem that cannot be
solved by XSLT 2. Is that correct?

On Wed, Dec 23, 2009 at 3:48 PM, David Carlisle <davidc@xxxxxxxxx> wrote:
>> I think I do need to do it. I cannot serve html as this would break my
>> company's accessibility guidelines (eg unclosed img tags)
> companies may have all kinds of guidelines, and sometimes thay have to
> be followed, the guidelines can be wrong though.
>> 1. I need to create a '<![CDATA[]]>' section in the output
>> ( explains why)
> the quoted page is misleading, it says:
>> This fixes the validator. The problem is that some older web browsers
>> don't understand the CDATA tag
> That implies that understanding <!CDATA[ markup is something fixed in
> newer browsers, but that is misrepresenting the situation. You can not
> have a CDATA section (or any markup) in a script element in HTML. The
> age of the browser is immaterial. Conversely if you generate XHTML and
> serve it as XHTML then CDATA quoting is understood by the browser.  So
> problems only occur if you generate XHTML and serve it as HTML. That
> generates syntax errors but commenting out odd lines only masks some of
> the problems, basically the whole document is then in error.
> Probably I haven't convinced you and you'll want to do this anyway, in
> which case, I'd use Martin's suggestion of writing the script to an
> external file.
> David

Current Thread