Subject: Re: Extension elements and namespace prefixes
From: James Clark <jjc@xxxxxxxxxx>
Date: Mon, 16 Aug 1999 12:28:09 +0700
|
XT doesn't support the extension element mechanism yet. The output
handler isn't where you're supposed to handle extension elements.
Even for output handlers, there is a bit of the problem here. An output
handler should not be stripping of namespace prefixes; rather if it
needs to recognize specific elements, it should be recognizing the
expanded name (the namespace URI/local name pair) not the qualified
name. At the moment do to this properly, it would have to process
namespace declarations, which is not very convenient (and a waste since
XT uses the expanded names internally, as any XSLT processor must). The
HTML output handler doesn't, which means it doesn't always do the right
thing. For example,
<xsl:stylesheet xmlns:xsl="http://www.w3.org/XSL/Transform/1.0">
<xsl:output method="html"/>
<xsl:template match="/">
<html>
<body>
<p>An html br element: <br/>.</p>
<p>Not at html br element: <br
xmlns="http://www.jclark.com/not-html"/></p>
</body>
</html>
</xsl:template>
</xsl:stylesheet>
will output
<html>
<body>
<p>An html br element: <br>.</p>
<p>Not at html br element: <br xmlns="http://www.jclark.com/not-html">
</p>
</body>
</html>
This is wrong because the second br element should not be recognized as
the HTML br element and so shouldn't get special treatment.
I'll fix this somehow; I haven't decided how yet.
Denys Duchier wrote:
>
> I have been playing with the new release of XT which I think is
> wonderful (thank you James). While adapting my own output handler to
> work with the new implementation I discovered a `problem' that had
> escaped my attention before:
>
> the name of the element passed to methods startElement and
> endElement contain the namespace prefix if it wasn't null.
>
> previously I always had a null prefix in such cases which is probably
> why I never noticed the problem. The reason for passing a prefixed
> name of course is that this is what is mandated by SAX. However it
> seems to me that this doesn't make sense for extension elements since
> the prefix is chosen by the stylesheet not by the implementation of
> the extension. Thus a robust implementation must strip off the prefix
> (btw, this is not done by XT's output handlers, which means that if
> they are invoked with an arbitrary prefix they don't work - at least
> the NXML output handler has this problem). At the moment, I simply
> look for a `:' in order to strip a prefix if any. My question is: is
> this the correct way to recognize a prefix? Is the namespace
> separator guaranteed to always be a colon?
XSL-List info and archive: http://www.mulberrytech.com/xsl/xsl-list
|
Cart
