Mark Struberg píše v Čt 04. 08. 2011 v 20:55 +0100:
> yes, replace javassist with BCEL and do that stuff ourselfs ;)
> There are a few other issues with javassist too. 

Sound good. Is javassist used only for "5.4. Client proxies" or for
something more?

> The good news: OWB is still much faster than all other DI containers 
> currently available :)
> 
Yes, I know and that's why I'm using it :)

> It will not be done before 1.2.0. We are currently working on getting an 
> owb-1.1.1 release done and then we can think about those issues.
> 
> But thanks for reporting this issue!
> 
> LieGrue,
> strub
> 
> --- On Thu, 8/4/11, Martin Koci <[email protected]> wrote:
> 
> > From: Martin Koci <[email protected]>
> > Subject: javassist RuntimeSupport.find2Methods contention
> > To: [email protected]
> > Date: Thursday, August 4, 2011, 7:49 PM
> > Hi,
> > 
> > 
> > I do some performance tests for hundreds concurrent users
> > and the
> > current winner of the 'most blocking methods hit parade' is
> > 
> > javassist.util.proxy.RuntimeSupport.find2Methods (used with
> > client
> > proxies) with this critical region:
> > 
> >      synchronized (methods) {
> >             if
> > (methods[index] == null) {
> >                
> > methods[index + 1] = thisMethod == null ? null
> >                
> >                
> >      : findMethod(self, thisMethod,
> > desc);
> >                
> > methods[index] = findSuperMethod(self, superMethod,
> > desc);
> >             }
> >         }
> >     }
> > 
> > But this is clearly javassist  internal code - it
> > there something that
> > OWB can do about it?
> > 
> > Regards,
> > 
> > Kočičák
> > 
> > 
> > 
> 


Reply via email to