forrest-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ross Gardler <rgard...@apache.org>
Subject Re: Catalogue application (Re: Dispatcher ClassCastException)
Date Fri, 23 Jun 2006 10:41:41 GMT
Brian M Dube wrote:
> Brian M Dube wrote:
> 
>> Ross Gardler wrote:
>>
>>> Brian M Dube wrote:
>>>
>>>> I'm building a site with 0.8-dev using dispatcher. I created a 
>>>> free-form (well-formed, but no DTD) XML file that describes the 
>>>> items for sale along with their options. I'm using the locationmap 
>>>> and sitemap to provide virtual URLs to view the items in the master 
>>>> file. 
>>>
>>>
>>> I'm just about to start experimenting with such a system. I wonder if 
>>> you would be able to open source your code, I'd be happy to turn it 
>>> into a plugin and assist with its development/maintenance since I 
>>> have a client with this need who is keen to open source the solution.
>>
>>
>> I can provide what I have done as open source. I need to finish the 
>> project as soon as possible, but the process of turning it into a 
>> plugin would provide early feedback about the design. How do you want 
>> to proceed?
> 
> 
> I have seeded a plugin locally. I have some questions. Does this master 
> file need to validate?

If there is a DTD defined in the document then it must validate.

> If so, does the plugin define the schema or does 
> the user?

Not sure what you mean here. Plugins can include a DTD's - they go in 
resources/schema. You need to reference them in the catalog.xcat file in 
the same directory. See the ListLocations plugin for an example.

> I really don't know which is easier for a future user of such 
> a plugin. I think the most flexible option would be internal support for 
> popular schemata with the option of extending the plugin with 
> user-defined schemata. Does this sound reasonable?

Yes, that is the approach I was planning to take. For my use case I need 
to pull the data from an OSCommerce back end. I was planing on 
converting it to a "popular schema" and then going from their into XDoc 
for Forrest (it's a static publish so performance is not a key issue).

I've done no research into which schema to use, so if you have one in 
mind I'll just go with that - no point in splitting our efforts.

Ross



Mime
View raw message