Re: [lists] [Springframework-developer] Expression language for use in ApplicationContext

[Colin S. <col...@ex...>]

Rod Johnson wrote:

>Peter,
>
>> richer expression language would have its use cases. The right thing to
>>    
>>
>do
>  
>
>>might be to make the language pluggable :^)
>>    
>>
>
>Yes, it could certainly be powerful. It has crossed my mind from time to
>time (except I didn't think of pluggability), but I was a bit concerned that
>the whole thing might get complex. However, I think it may be worthwhile.
>  
>
I've used OGNL with Tapestry, and I find it quite useful and powerful. I 
know WebWork2 also uses it now. While pluggability isn't a bad idea, I 
would be happy even with just OGNL... W/regards to the complexity 
aspect, where do you see issues? One area that's a bit grey relates to 
the lifecycles of the other beans...

>  
>
>>On a not entirely unrelated note, one of the things I've been missing from
>>Spring is what you might call "anonymous beans". In some situations, I
>>    
>>
>don't
>  
>
>>really want to pollute the namespace with beans that will only ever be
>>    
>>
>used
>  
>
>>in one place:
>>
>>   <bean id="foo" class="eg.Foo">
>>    <property name="bar"><ref bean="bar"/></property>
>>   </bean>
>>
>>   <bean id="bar" class="eg.Bar"/>
>>
>>But would like to be able to write something like
>>
>>   <bean id="foo" class="eg.Foo">
>>    <property name="bar"><bean class="eg.Bar"/></property>
>>   </bean>
>>
>>Has this been discussed before?
>>    
>>
>
>I don't think it's been discussed on the list, but I've considered the idea
>of reducing bean visibility. Nesting this way would be one way to do it
>(although it might involve substantial refactoring). Another way I've
>considered is the notion of "protected" and/or "private" beans, that can't
>be accessed outside that factory, but can be used in references. Certainly I
>think agree in some cases it would be good to avoid filling the namespace
>with single-use beans.
>
>What do others think?
>  
>
Actually just the other day I was wondering why beans couldn't be 
defined inline like this. I think it's a lot cleaner when you want to do 
one-off bean definitions...

>I suspect that this is a 1.1 thing, as I think we need to focus on existing
>functionality for 1.0. But it's a good suggestion nonetheless.
>  
>