Before we jump the gun on this I'd like to see if more
people could give some 
feedback  on the pros and cons of this.  I definitely
some advantages in that it 
simplifies the code some and is based on technologies
that more developers are 
more likely to be familiar with.  I do think that
something is lost in that SDO 
's would provide greater runtime referential checking
with the specification. 
I'm not in any what advocating not taking this route
just hoping before we make 
such a leap that a more livelier discussion would take
place.

ant elder wrote:
> I like this a lot. Lets change everything over to it
from SDO.
> 
>    ...ant
> 
> On 3/14/06, Jeremy Boynes <[EMAIL PROTECTED]>
wrote:
>> I have got the StAX stuff to the point where I can
run all the itests in
>> the build and the tomcat/testing tests using the
StAX framework. I think
>> this is an opportune time to open discussion on
whether we should switch
>> over to this once and for all.
>>
>> You can enable the framework by setting the useStax
boolean in
>> BootstrapHelper to true and rebuilding the core
module (and any others
>> you wish to test).
>>
>> To see what is involved in adding an extension have
a look at:
>> JavaScriptImplementationLoader for the .js
container
>> JSONRPCBindingLoader for the .jsonrpc binding
>>
>> Comments welcome...
>> --
>> Jeremy
>>
>> Jeremy Boynes wrote:
>>> I checked in a framework for a StAX-based
configuration loader for the
>>> SCA core. It is based on a set of element handlers
that generate a model
>>> object from a element in the XML stream; handlers
for the core and
>>> system schemas are in the core module, handlers
for extensions can be
>>> bundled in the extension module and are
contributed as regular system
>>> components.
>>>
>>> I tested with the basic system and helloworld
configurations and
>>> discovered a problem with the way configured
properties and references
>>> are handled. This should be resolved once the
FIXME in ComponentImpl is
>>> addressed.
>>>
>>> To get this working I had to modify the autowire
resolver in the
>>> SystemAggregateContextImpl. I hope this can be
removed once Jim gets
>>> references working. I also think we need to
reexamine the autowire
>>> algorithm but will open another thread for that.
>>>
>>> --
>>> Jeremy
>>
> 

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

Reply via email to