RE: [aspectj-users] Serial Version UID

[hilsdale@xxxxxxxx]

Vincenz,

This seems at first glance to be a reasonable idea.  I know that I, at least, haven't looked back into serialVersionUID since we started working with bytecodes.  This would be a great thing to add to the bug database as a feature request.

I don't think it will get into 1.1, though, since we've released the release candidate and as such are squashing bugs rather than adding features.

-erik

-----Original Message-----
From: Vincenz Braun [mailto:vb@xxxxxxxxx] 
Sent: Thursday, 10 April, 2003 10:01 am
To: aspectj-users@xxxxxxxxxxx
Subject: [aspectj-users] Serial Version UID


Hello,

the FAQ states:

"However, changes to the serialVersionUID attribute are typically only
important when using serialization for the long-term persistence of objects.
Using standard Java serialization for long-term persistence has a number of
drawbacks and many developers already use alternative solutions. For one
possibly standard solution, see Lon g-Term Persistence for JavaBeans
Specification."

I think that is only party true. This issue is really important for client -
server applications
depending on object serialization (RMI, EJB etc.). I would not call it
long-term persistence. Introducing
aspects on client or server side without breaking serial version UID is very
important in my opinion. 
Now with using byte-code weaving would'nt it be easy to provide a feature
that can handle serialVersionUID
a better way? For example having a flag that introduces a serialVersionUID
field with the value of original 
class (Either with warnings if there were introductions or ommitting such
classes)?

It would be great if such a thing could be integrated into the final 1.1
release.

Thanks,
Vincenz

_______________________________________________
aspectj-users mailing list
aspectj-users@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/aspectj-users