Hi Torsten, Thanks for your reply.
> And what about the root tag? > You want to keep the pipeline well-formed. Well Cocoon does not garantuee this. It is a common mistake (did it myself quite often) to forget in a XSLT stylesheet to always set a root tag (<template match="/"><myroot>...). > Cocoon is *streaming* the XML through the pipelines. > That's what you would expect from a today's XML pipeline > architecture ...which is also the reason why you cannot > react on XML values inside the stream from the sitemap > level without buffering the whole process. I was thinking out-of-the-box about functional/conceptual requirements. Your reply is about technical implemention/performance issues. Of course these are sometimes hard to separate because you cannot build something which is impossible (my computer still makes no breakfast). But really, if you take a step back it is hard to explain to new XML pipeline users why they can select on a request parameter but not on a XML field. Regards, Kees -----Oorspronkelijk bericht----- Van: Torsten Curdt [mailto:[EMAIL PROTECTED] Verzonden: Wednesday, August 10, 2005 12:36 PM Aan: dev@cocoon.apache.org Onderwerp: Re: Architectural concerns > Now to me the first generator is already a farce. I made an SQL > transformer that simply accepts the query as parameter in stead of > using > the XML stream. I always thought it would be better to have a sql query being executed *before* the transformation starts. So from that point of view this does makes sense. Altough... having sql statements inside the sitemap! *shiver* > So this means we should allow: > - not to have a generator at all (yes I thought about the > consequences) ? <snip/> > - generator: starts a new stream (thus ignoring previous stream if > present) > - transformer: changes a stream (which could be empty at the > beginning) > - serializer: ends the stream And what about the root tag? You want to keep the pipeline well-formed. Sorry, not excited here. > Related to this issue is the Cocoon Aggregate statement Have a look into the cinclude transformer <snip/> > 2. selectors are evaluated before the pipeline is executed > > This basically means you can select on global, request and session > fields because these are already set. But of course you also want to > react on XML values in the stream. That is what you expect from an XML > pipeline architecture. Cocoon is *streaming* the XML through the pipelines. That's what you would expect from a today's XML pipeline architecture ...which is also the reason why you cannot react on XML values inside the stream from the sitemap level without buffering the whole process. > In short: I want to be able to choose different transformers depending > on the elements/attributes placed in the stream by the previous > transformer. Maybe you want to have a look into cocoon 1 :-P cheers -- Torsten
