Stefan Bodewig wrote:
On Tue, 20 Apr 2004, Matt Benson <[EMAIL PROTECTED]> wrote:
I'd like to go ahead and merge it to the 1.6 branch...
+1
before that happens we should hopefully resolve any
conflict regarding the name of it... I personally am
glad enough to have it under any name... so any
conflict is (so far) between you and Stefan on that
issue...
Conflict is too big a word.
Thank you for bringing this up again, I have not forgotten about it
(nor about reviewing the redirector stuff).
Let's see. ContainerMapper allows us to use more than one mapper and
it knows about two modes of operation. In the first mode, the
contained mappers are chained to each other so that mapper two gets
the results of mapper one and works on it. In the second mode, each
mapper gets the original from and the result is the union of all
results of the contained mappers. Is this correct?
Yes
I'd probably prefer this implemented by two classes and call them
CompositeMapper and ChainedMapper (MapperChain?) - probably using
a ContainerMapper (MapperContainer?) as abstract base.
This makes sense, It is I think Matts original idea. The only
thing to keep in mind is that the following (I think) should
still work.
<copy ...>
<mapper chain="yes">
<globmapper .../>
<mymapper ../>
</mapper>
</copy>
as against:
<copy ....>
<mapper>
<chainmapper>
...
</chainmapper>
</mapper>
</copy>
or
<copy .....>
<chainmapper>
...
</chainmapper>
</copy>
BTW, it would probably be a good idea to remove duplicates in
non-chained mode, at least as an option. Using a Set instead of a
List for ret would do that.
This shoulds good, and is a good reason to split up the containerclass.
Peter
Stefan
---------------------------------------------------------------------
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]
