Re: MethodClosure (BCEL Delegator implementation)

Leo Sutic wrote:
> 
>  > The commons people are busy putting dependancies on CommonsLogging
>  > into things. Its cursing AltRMI now because of a dependecy on
>  > Attributes and HttpClient. Should we really be a in a rush to
>  > donate components?
> 
> Paul,
> 
> when we donate comonents there is an implicit trust that the receiving 
> project will evolve the code in the spirit of the original and keep in 
> mind that the code just received is a dependency for some other project. 
> That's the downside of donating code. The upside is that unless we do 
> donate code, the scope of Avalon will start growing. The dependencies 
> created can also be seen as something good, as it forces cooperation 
> among Apache projects.

Yup, well said.

> That said, I'm worried about the inclusion of Commons logging into 
> everything, as (if I understand it correctly) it isn't IoC, thus forcing 
> the whole system into using Commons logging. Could you be more specific 
> about in what way AltRMI is suffering? If we're going to talk to the 
> Commons Logging people to see if a compromise can be reached (and I 
> think we should), it would help to have a list of issues.

There was a thread that GeirMJ started some months back on commons-dev, 
because he wanted to use Commons Logging with an IOC method.

Anyway, since it's "commons", we sould really see that some compromise 
can be reached, even if I actually don't see the technical way... if a 
class gets the logger via a static method... unless we can invert the 
deliver on the *package*, and the static method gets it there, per 
instance. Just random thoughts...

-- 
Nicola Ken Barozzi                   [email protected]
             - verba volant, scripta manent -
    (discussions get forgotten, just code remains)
---------------------------------------------------------------------