"Daniel Fagerstrom" <[EMAIL PROTECTED]> wrote:With XSLT and agreagate you could write something like:
<map:match pattern="/news/*"> <map:aggregate element="page"/> <map:part src="cocoon:/data/articles/{1}"/> <map:part src="cocoon:/data/authors/{request-param:author}/author"/> <map:part src="cocoon:/data/lists/topnews"/> </map:aggregate> <map:transform src="/templates/news.xslt"/> <map:serialize/> </map:match>
No, I can't... You see where do you get the author from? You get it from a request parameter. While in my case the author information (or ID) is stored in the article original XML data... It would be fair enough if I could rewrite it like
<map:match pattern="/news/*">
<map:aggregate element="page"/>
<map:part src="cocoon:/data/articles/{1}"/>
<map:part src="cocoon:/data/authors/{document(cocoon:/data/articles/{1})/article/autho
r}"/>
<map:part src="cocoon:/data/lists/topnews"/>
</map:aggregate>
<map:transform src="/templates/news.xslt"/>
<map:serialize/>
</map:match>
But, darn, that's UGLY! :-(
Pier
Ok, I missed that. A new trial:
<map:match pattern="/news/*">
<map:aggregate element="page"/>
<map:part src="cocoon:/data/articles/{1}"/>
<map:part src="cocoon:/data/lists/topnews"/>
</map:aggregate>
<map:transform src="/templates/news.xslt"/>
<map:serialize/>
</map:match><HTML xmlns:xsl="http://www.w3.org/1999/XSL/Transform"; version="1.0">
<HEAD>
<TITLE><xsl:value-of select="/page/article/title"/></TITLE>
</HEAD>
<BODY>
<H1><xsl:value-of select="/page/article/title"/></H1>
<P>
<xsl:variable name="author" select="document(cocoon:/data/authors/{/page/article/@author}/author)"/>
Author:
<A href="mailto:{$author/email}";>
<xsl:value-of select="$author/name"/>
</A>
</P>
<xsl:copy-of select="/page/article/body"/>
<DL>
<DT>Other top news from us</DT>
<xsl:for-each select="/page/related/articleref">
<DL>
<A href="/news/[EMAIL PROTECTED]"><xsl:value-of select="title"/></A>
</DL>
</xsl:for-each>
</BODY>
</HTML>
This is partly pull driven and quite close to your original template, and in a non-xml syntax it would of course look a litle bit nicer. If you use the document() function with run time dependent data you can forget about cashing, but the same applies to your original formulation. If you want cashing I'm afraid that you have to do something like what you proposed above.
/Daniel Fagerstrom
