I added PredicateDecorator marker interface.
Stephen

----- Original Message ----- 
From: "Sandor Murakozi" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, May 19, 2004 10:49 AM
Subject: Re: [collections] [PATCH] Commons Collections - Wrapper and 


> Hi Stephen,
> 
> I also think a kind marker interface would be nice. If
> someone wants to 
> travel a tree of predicates (which is needed to display it's
> structure, save it, almost anything except building and
> simply invoking it) then it would be extremely useful. It
> would be also handy if you want to handle different types of
> Predicates differently. I think a great part of Predicate
> users will need some similar functionality (including me in
> my actual project :-)). Now we have to create our reflection
> based solution, or huge if else-if statements (brrrrr) , so
> I would vote on marker interfaces.
> 
> And yes, it would probably open the door, but is it really a
> problem? (I use only predicates at this time, so I'm not a
> transformer expert, but I can imagine transformer users with
> the same problem then me).
> 
> BR,
> don
>  
> >Thank you for your patch.
> 
> >I'm not certain as to how to procede on this one. There is
> nothing wrong
> >with your patch or idea. The problem is that it opens the
> door to accessing
> >other functors such as Transformer and Closure. How far
> does this go?
> 
> >One idea I have considered is to have a single generic
> Decorator interface
> >that specifies one method which returns a predicate.
> 
> >Another idea is to use reflection, with a single method on
> PredicateUtils to
> >access this functionality. This might be less intrusive,
> but perhaps is also
> >less obvious.
> 
> >Not sure, any opinions?
> 
> >Stephen
> 
> >----- Original Message -----
> >From: "Brian Lloyd-Newberry" <[EMAIL PROTECTED]>
> >To: <[EMAIL PROTECTED]>
> >Sent: Friday, May 14, 2004 6:38 PM
> >Subject: [PATCH] Commons Collections - Wrapper and
> Composite >Predicates
> 
> 
> > Commons Collections developers,
> >
> >    In looking at the 3.1-dev source I see that some
> implementations of
> > the Predicate objects have:
> >    Predicate getPredicate()
> >    or
> >    Predicate[] getPredicates()
> > methods. This is a good thing for me, as I am developing a
> search
> > interface and would like to be able to use predicates in the
> > search/filter process, and be able to do a display of the
> predicate
> > expression structure.
> >
> >    I observe that there are basic types of predicates.
> Ones which
> > operate on Objects, ones that operate on a single
> (sub)Predicate and
> > ones which operate on more than one Predicate. I propose
> that the single
> > and multiple cases be tagged using an interface so that
> the appropriate
> > getPredicate() or getPredicates() method can be found.
> >
> >    I have attached a patch and 2 new interfaces
> (WrapperPredicate.java
> > and CompositePredicate.java) which provide the
> functionality I propose.
> > WrapperPredicate declares Predicate getPredicate().
> CompositePredicate
> > declares Iterator predicates(), shielding the
> CompositePredicate from
> > modification (the Iterator returned is Unmodifiable) that the
> > Predicate[] getPredicates[] implementation would allow
> unless a
> > defensive copy was made before the return took place.
> >
> >    If you have any questions or concerns about my
> implementation please
> > let me know. This is my first patch on a Jakarta project,
> so I hope I
> > have  done everything correctly (cvs diff -u).
> >
> >    Thank you for your work and for providing such a useful
> library of
> > collections to the community.
> >
> > -Brian
> >
> > Brian S. Lloyd-Newberry
> > Software Engineer
> > RBS, Inc
> > http://www.rbs2000.com
> >
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]


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

Reply via email to