Re: [picocontainer-dev] Pico2: ASM ImplementationHiding ComponentAdapter

Jörg,

This to me personifies the issue I had with the old solution :-

     http://svn.picocontainer.codehaus.org/browse/picocontainer/java/ 
sandbox/baby-steps/nano2/container/src/test/org/nanocontainer/script/ 
xml/issues/Issue0170TestCase.java?r1=3351&r2=3373

A component itself, or any components depending on it, could replace  
the implementation if they were cunning enough to find the right  
method.  That should be a privilege for containers only.

- Paul


On May 13, 2007, at 10:18 PM, Jörg Schaible wrote:

> Hi Paul,
>
> Paul Hammant wrote on Sunday, May 13, 2007 6:53 PM:
>
>> I've written one that does simple ImplementationHiding and changed
>> the HotSwappable one to extend that.  It brings us right up to date
>> with ASM.  Previously we used to use ProxyToys 0.21 which Used Cglib
>> 2.0.3 (the latest despite being dated 2005) which bundled (by jarjar)
>> and ASM release in the 2.x series from then.
>>
>> Gems also has a poolable component that's much more advanced.  I
>> think it might be nice to upgrade that to use ASM direct too.
>> Thoughts?
>>
>> Maybe we could go as far as jarjar-ing the bits of ASM into
>> out jar.
>> ASM is 42K.
>>
>> Thoughts?
>
> I think you have to! cglib bundles ASM 1.x and this was necessary  
> since ASM 1.x and 2.x was not compatible. I don't know how  
> compatible ASM 2.x and 3.x is though. OTOH I don't know, why you  
> dropped proxytoys. It was mainly refactored code of pico with an  
> abstraction layer. Adding direct ASM support to ptoys would have  
> been quite easy.
>
> - Jörg
>
> ---------------------------------------------------------------------
> To unsubscribe from this list please visit:
>
>     http://xircles.codehaus.org/manage_email
>
>


---------------------------------------------------------------------
To unsubscribe from this list please visit:

    http://xircles.codehaus.org/manage_email