Re: Bug in 0.10-RC2?

It is calculated based on what Javassist provides. A snip hacked by Bob Lee
that worked well so far, based on the specs (depends on methods etc).
It is calculated at the weaving time, based on the initial bytecode (before
any transformation).
I ll check the source and check the test coverage for it.

Thanks
Alex

----- Original Message -----
From: "eu" <eu-E/[email protected]>
To: "Alex" <alex-he/e8N7AIe/[email protected]>
Cc: "aspectwerkz-devel" <aspectwerkz-devel-81qHHgoATdGxIXFVlbCvtR2eb7JE58TQ@public.gmane.org>
Sent: Friday, May 14, 2004 8:03 PM
Subject: Re: [aspectwerkz-devel] Bug in 0.10-RC2?


> Alex wrote:
>
> >Could you post your pointcut defintions ?
> >There is a triggering when a class gets weaved so that we add the
> >serialVersionUID at the end of the transformation chain, but it seems
that
> >in your snip this does not gets triggered as it should.
> >
> >
>   How it is calculated?
>
>  I usually use something like this:
>
>  ObjectStreamClass objStreamClass = ObjectStreamClass.lookup(class1);
>  long id = objStreamClass.getSerialVersionUID();
>
>  Unfortunately it assumes that class must be loaded but it is the most
> safe way having in mind all of those javac, gjc, Eclipse and so on
> compillers.
>
>   Anyway, here is my definition.
>
> <aspectwerkz>
>   <system id="test">
>     <include package="com.cibcwm.go.otis.dasl"/>
>     <exclude package="com.cibcwm.go.otis.dasl.console"/>
>     <exclude package="com.cibcwm.go.otis.dasl.cmp"/>
>     <aspect class="com.cibcwm.go.otis.aspect.ThreadCaptureAspect"
> deployment-model="perThread">
>       <pointcut name="pc1" expression="execution(*
> com.cibcwm.go.otis.dasl..*.*(..))"/>
>       <pointcut name="pc3" expression="call(*->* com.capco..*.*(..))"/>
>       <pointcut name="pc4" expression="call(*->* java.sql..*.*(..))"/>
>       <advice name="capture" type="around" bind-to="pc1"/>
>     </aspect>
>   </system>
> </aspectwerkz>
>
>     Eugene
>
>
> >----- Original Message -----
> >From: "eu" <eu-E/[email protected]>
> >To: "aspectwerkz-devel" <aspectwerkz-devel-81qHHgoATdGxIXFVlbCvtR2eb7JE58TQ@public.gmane.org>
> >Sent: Friday, May 14, 2004 5:41 PM
> >Subject: [aspectwerkz-devel] Bug in 0.10-RC2?
> >
> >
> >
> >
> >>Hi,
> >>
> >>  It seems that there are some small issue with serialization. If AW is
> >>instrumenting Serializable classes in the runtime and serialVersionUID
> >>is not explicitly declared, then client app, which may be using not
> >>instrumented classes will get a error. I think, that instrumentation
> >>should insert an explicit serialVersionUID that match with the original
> >>class.
> >>
> >>  regards,
> >>  Eugene
> >>
> >>
> >>_______________________________________________
> >>aspectwerkz-devel mailing list
> >>aspectwerkz-devel-81qHHgoATdGxIXFVlbCvtR2eb7JE58TQ@public.gmane.org
> >>http://lists.codehaus.org/mailman/listinfo/aspectwerkz-devel
> >>
> >>
> >>
> >>
> >
> >_______________________________________________
> >aspectwerkz-devel mailing list
> >aspectwerkz-devel-81qHHgoATdGxIXFVlbCvtR2eb7JE58TQ@public.gmane.org
> >http://lists.codehaus.org/mailman/listinfo/aspectwerkz-devel
> >
> >
>
>