Paul Hammant wrote:
> When you say loop, that never escapes, do you mean that becase of the
> lack of exception? I will commit the patch you supply.
I can commit if I have a patch. Just don't have enough of a grasp on
the low level altrmi code to fix this easily :-)
> The 'missin beanimpl xx' problem is likely to be caused by a
> handshaking issue between client and server. I'll take a look at it.
> Does the exception ever occur if the JVM has not been suspended? Do
> you mean like suspend-laptop in MS Windows? .. the whole OS is dumped
> to the hard drive?
I have seen this and the second problem at the bottom of the message
when the server-side JVM is restarted as well. The client seems to be
reconnecting, but not realizing that the JVM is different from the one
it was just connected to.
Yup, one of my machines is an XP notebook that can be suspended to ram
or disk. Most things resume without any problems. But sometimes
timeouts will fire because the system clock appears to jump ahead by
several hours.
In the following code in DefaultMethodInvocationHandler.java, there are
cases where
wr is null, and cases where wr is not null, but beanImpl is. I added
the throw in the
case where wr is null to avoid the NPE in the next line of code. This
does not seem to
be what is needed to tell the client to stop trying to reconnect with
the same information
over and over and over again.
---
WeakReference wr = (WeakReference)mRefBeans.get(
request.getReferenceID() );
if( wr == null )
{
System.out.println( "missing weak reference for
referenceID - "
+ request.getReferenceID() + " pub "
+ mPublishedThing );
return new NoSuchReferenceReply( request.getReferenceID() );
}
Object beanImpl = wr.get();
if( beanImpl == null )
{
System.out.println( "Missin beanImpl for " +
request.getReferenceID() );
return new NoSuchReferenceReply( request.getReferenceID() );
}
---
Cheers,
Leif
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
