On Thursday 27 June 2002 06:40 am, Stefano Mazzocchi wrote:
> Counterexamples:
>
> 1) FileGenerator, where the schema of the output is bound to the
> resource being parsed and it's available at runtime only.

But what if there was something like:

<map:generate type="file" src="cocoon://myfile.xml">
  <map:output-schema src="cocoon://myfile.xsd"/>
</map:generate>

> 2) XSLTTransformer, where the transforming instructions (thus the schema
> constraints) are bound to the stylesheets and not the code that
> interprets it. (Things change with XSLTC which is able to compile
> stylesheets, but XSLTC will never be Avalon-aware anyway).

And:

<map:transform type="xslt" src="cocoon://mytransform.xsl">
  <map:input-schema src="cocoon://mytransform-input.xsd"/>
  <map:output-schema src="cocoon://mytransform-output.xsd"/>
</map:transform>

Upon further dwelling, doing this at the pipeline level would probably be 
overkill, but for the blocks you have been working on, it might make sense at 
the block entry/exit points.

> Still, I have the feeling that Cocoon should not use the same container
> to handle sitemap components and the other more general components.

I agree. Cocoon's sitemap is a special case and will require a specialized 
container.
-pete

-- 
peter royal -> [EMAIL PROTECTED]

--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to