|
Sort Order:
We would, as a minimum, need to hook into the CGLIB ASM usage to change this. I can't forsee a reason why this would be a problem but, the devil is in the details... Aren't the CGLib proxies themselves final? I got this from different double-proxying threads in the forum where the second (mostly unwanted) proxying fails due to unextendable "classes" of the first proxying when the result of the first one are CGLib proxies. See for example http://forum.springframework.org/showthread.php?t=40829#3 Jörg That is the case, but we may well be able to change that as needed. It's not currently on the roadmap to do anything here. We might eventually do a CGLIB alternative for class proxying, but that's still a while away. Juergen |
||||||||||||||||||||||||||||||||||||||||||
Improvement
Resolved
Major
please don't generate a proxy override methods equals() and hashCode() using "final"
CGLIB seems to make those methods final by default - and there is no obvious option to keep them non-final.
Rob, do you have any insights to share here?
Juergen