On Sun, 5 Aug 2001, robert burrell donkin wrote:

> On Sunday, August 5, 2001, at 12:53 AM, Craig R. McClanahan wrote:
> 
> > On Mon, 30 Jul 2001, robert burrell donkin wrote:
> >
> >>
> >> so i guess IMHO digester would benefit from a flexible mechanism which
> >> would give users the choice from a set of simple - but easy to learn -
> >> rules which are based on the current format and a sophisticated - but 
> >> more
> >> complex - set of matching rules that work using some variation of XPATH.
> >>
> >> the best of both worlds? just maybe :)
> >>
> >
> > Just checked in the mechanism to make this easy.  If you create a cool
> > Rules implementation that implements some other policy, it can certainly
> > be made available as part of the standard Digester package.
> 
> excellent :)
> 
> i've just had a quick look at your enhancements and they really pretty 
> good.
> 

Cool.

> sorry if i was a bit pushy earlier.

Not a problem.

> it's just that's i'd really, really 
> like to use digester (which is better designed and more sophisticated) to 
> replace a stack-based parser in some generation tools created to help 
> refactoring in jakarta-ecs. i bootstrap by generating against the 
> xml-schema schema so i need digester to be able to handle big, complex 
> schema. since this material will probably end up in the commons (quite 
> possibly in a project as outlined by jason earlier) i'd much rather see 
> enhancement classes in digester where they belong.
> 

Sounds like a good reason to enhance Digester to suit the needs.  I just
don't want to violate our "backwards compatibility" principles (in the
Commons charter) for existing Digester users.  Extensibility lets us have
our cake and eat it too.

> i'm feeling pretty confident at the moment that everything's going to work 
> out well :)
> 

If you find yourself creating one or more Rules implementations that you
think make sense to include in the standard Digester package, feel free to
throw 'em this way.  That might save other people from having to solve the
same problem again.

And it wouldn't hurt my feelings at all if some kind folks added some good
JUnit test cases for Digester :-).

>
> - robert
> 

Craig

Reply via email to