cool!! :), when I looked at the code yesterday (when jaliya
mentioned the problem) I saw it and thought it is better to discuss it
in the list ... so other deveopers get to know about the stateless
assumptions we made.
Thanks
Srinath
On 3/22/06, Eran Chinthaka <[EMAIL PROTECTED]> wrote:
> You being one of the pioneers in Axis2, how can you talk abt some old
> code ? ;-) Get an svn up and you will see they are no more like that.
>
> Srinath Perera wrote:
> >>> On 3/21/06, Eran Chinthaka <[EMAIL PROTECTED]> wrote:
> >>>
> >>>> I have not kept this attribute across multiple invocations of the same
> >>>> handler. Its *within* one invocation of that handler.
> >>>>
> >> Are you keeping an instance var that holds some value across two or more
> >> methods in the same handler? That is, do you use an instance var to pass
> >> some state from the main method to other methods? If so that will lead
> >> to problems and can easily be fixed with params.
> >>
> >> If not I'm confused what the offending thing is; can someone point out
> >> the code please.
> >>
> > the class is org.apache.axis2.handlers.addressing.AddressingOutHandler,
> >
> > offending part is (following is yesterdays code) Handler attributes
> > addressingNamespaceObject and msgCtxt .. they remember things between
> > the method invocations (addressingNamespaceObject accsessed by most
> > method without passing it in .. e.g.
> > processToEPRReferenceInformation(....)) and run in to trouble
> >
> > anothor thought .. we should be careful when hanging to the
> > messagecontext from some instance varaible as that might lead to
> > memeory leaks if object (e.g. Handler) is never garbage collected
> >
> > Thanks
> > Srinath
> >
> > public class AddressingOutHandler extends AddressingHandler {
> > ..............
> > protected OMNamespace addressingNamespaceObject;
> > private MessageContext msgCtxt;
> > public void invoke(MessageContext msgContext) throws AxisFault {
> > .....................
> > }
> > }
> >
> > --
> > ============================
> > Srinath Perera:
> > http://www.cs.indiana.edu/~hperera/
> > http://www.bloglines.com/blog/hemapani
> >
> >
>
>
--
============================
Srinath Perera:
http://www.cs.indiana.edu/~hperera/
http://www.bloglines.com/blog/hemapani
