On Fri, 10 Sep 2004 08:57:19 +0200, "Eric van der Vlist" <[EMAIL PROTECTED]> said: > ... The other thing I find misleading and probably under documented is > the relation between the Page Flow Controller and XPL pipelines. > > Because I was already familiar with Cocoon when I started playing with > Presentation Server, I made the assumption that everything would be > defined in pipelines and it took me a while to understand that the > Page Flow Controller is a black box that acts as kind of default > pipeline and you might want to make that more explicit in the > documentation. > > I assume that there are many good historical reasons why that's like > this, but I think that it would be cleaner to consider that the PFC is > an actual pipleline and to show that in the page-flow.xml files and to > use XPL directly in these files. > > Instead of: > > <page path-info="/login" model="login/model.xpl" > view="login/view.xml"/> > > I think that: > > <page path-info="/login"> > <p:processor name="oxf:PageFlowControler"> > <p:input name="model" href="login/model.xpl"/> > <p:input name="view" href="login/view.xml"/> </p:processor> </page> > > That's more verbose but that shows well the nature of the PFC and also > that would let people put XPL directly in page-flow.xml documents.
First let me throw out the following dislclaimer : I am not the author of the PFC and I have only been working around this stuff for a few months. That said I have a mixed reaction to the above. On the one hand I agree that there are some things that could be conveyed more clearly by PFC file's content. On the other hand I think that allowing XPL in the PFC file would be at odds with the goal separating out the concerns of page logic, page flow, etc. ( Am envisioning people creating monolithic PFC files and then getting lost in the content. ). In fact I am not sure I like the fact that one can use xupdate in the PFC file. Couple of final items that may be of interest. First the PFC is itself a processor. You can see it being used in displatch.xpl in the examples. Secondly, while the main processor is typically the PFC you could use another processor, e.g. the pipeline processor, as the main processor. Now perhaps this is off topic or of only acedamic interest, but this does mean that you could potentially generate your page flow from anything. i.e. A file that looks like what you have proposed. Attached is a trivial sample that shows : o Using the pipe line processor as the main processor. o Usage of the page-flow processor from within XPL. o Getting the page request info. I suppose if you wanted to go 'whole hog' you could ommit the use of the page flow processor altogether and just use the pipe line processor. Not sure what the consequences of that would be. That is other than forcing you to type a lot more. -- Regards, Dan S
some-main.not-a-zip-honest
Description: Zip compressed data
