Well ... i've added a few log traces, and, well.. this is getting me
nuts ...
What i can see, is that, sometimes, the msgContext attribute from the
Call object has a different hashCode () between the constructor method
and the invoke() call...
Even more weird, i can the the same msgContext hashCode() (means : "same
objects" since hasCode is not overridden, AFAIK) shared among different
Call objects ...
So, either, there is something really wrong with my JVM, thread
handling, etc, or there is something i don't catch ...
Gonna try to update my JVM and see if it changes anything...
Carpe Sebastien wrote:
makarand sovani wrote:
Are you using wss4j ? We have recently come across non-thread safe
behaviour in wss4j.
If you are not using wss4j can you check your handlers to see if they
are threadsafe ?
Makarand
Well, i'm not sure about what you're pointing me to ...
my client code for performing call, is as simple as
Service s=new Service();
Call c=(Call)s.createCall();
c.setMaintainSession(true);
// setting parameters/returnType/parameters
....
c.invoke(parameters);
No special handler what so ever... No config file to change default
behavior.
Am i using wss4j ? well, no... I don't think so (the closest stuff i
got in my classpath is wsdl4j, but an 's' and a 'l' doesn't make a s,
do they ;) ) ...