Rich,

I really need some time to review this throughly, on the surface it
seems like a way to get around the deliberately designed architecture,
more specifically to be able to completely bypass AXIOM. Is that the
design criteria for this? If so, let's state it explicitly and take a
VOTE on it. Either way, i personally need a few more days to spend
some time on this. RC2 is completely sucking up my waking hours. Can
we please do this after a few days? a week?

thanks,
dims

PS: /me dreading the branch merge nightmare.



On 10/30/06, R J Scheuerle Jr <[EMAIL PROTECTED]> wrote:


Yes, this is a problem.  Axis v1.0 had the following on its MessageContext:
 static ThreadLocal currentThreadMessageContext = new ThreadLocal();

 So in order to implement the solution, we would need to add this capability
to the Axis2 MessageContext
 and set the currentThreadMessageContext in the transport utilities when the
Message is received.

 Dims/Sanjiva,
 Would you allow us to change MessageContext and the transport utils to set
the MessageContext on TLS ?



 >FWIW, on the 1.1 branch,
 >MessageContext.setCurrentMessageContext(msgContext);
appears to only
 >be called from AbstractMessageReceiver which is called after the
 >handler chain.

 >David


 On 30/10/06, R J Scheuerle Jr <[EMAIL PROTECTED]> wrote:
 >
 >
 > Okay, I see your point here.
 >
 >  So I believe your suggestion is for the user to provide and configure a
 > javax.xml.stream.XMLInputFactory (outside of Axis2).
 >  The XMLStreamReader provided by this XMLInputFactory would create and
could
 > act as a Filter of an actual (woodstox) XMLStreamReader.
 >  The user-provided XMLStreamReader could implement BuilderAwareReader,
 > giving it access to OM Elements.
 >  The user-provided XMLStreamReader could obtain the Axis2 MessageContext
 > from thread local storage.
 >
 >  We will investigate this solution.
 >
 >  Followup Question:
 >  If we decide that we can do this outside of Axis2, could we commit a
test
 > to Axis 2 to ensure:
 >     a) The BuilderAwareReader works
 >     b) The Axis 2 MessageContext remains available on thread local
storage.
 >
 >  The test could be useful to other Axis 2 users that have the same
concerns.
 >
 >  On 10/26/06, Sanjiva Weerawarana <[EMAIL PROTECTED]> wrote:
 >  > On Thu, 2006-10-26 at 20:50 -0400, Davanum Srinivas wrote:
 >  > > I completely agree that this is mixing layers of processing and the
 >  > > stax filter should not have access to it. Also, at the risk of
stating
 >  > > the obvious...We don't need to pass in the MessageContext at all. In
 >  > > current SVN, you can get the message context using the thread local.
 >  >
 >  > Except that while the StAX code is invoked there may not be not be one
 >  > around at all.
 >  >
 >  > Sanjiva.
 >  >
 >  >

 Rich Scheuerle
 IBM Web Services
 Apache Axis2 ([EMAIL PROTECTED])
 512-838-5115 (IBM TL 678-5115)




--
Davanum Srinivas : http://www.wso2.net (Oxygen for Web Service Developers)

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to