On Thu, 2007-08-16 at 11:23 +0100, Darran Lofthouse wrote: > Thomas, > > Would it be possible for you to have a look at this - if this is > something we should be supporting I can get a test case together and get > the fix upstream. > > Sujeet has confirmed that they need to configure the handler at the > endpoint level and not globally in the pre-handler-chain so the > questions really is should it be possible to use a handler defined for > the endpoint that transforms a message so it is valid for the > 'unbindRequestMessage' method? Yes, a handler should be able to transform the DOM view of the incoming message. If done so, the message is marked as modified and will be run through the unbind algorithm again. It should not be necessary to modify the handler processing logic, since this is a common use case that we have test coverage for. > > The fix that went into the patch to defer the call to > 'unbindRequestMessage' until after the handlers is only causing a > regression in one test case so it could be that that is an invalid test > that needs fixing. >