> I'd be willing to help with this issue, but from my point of view there
are
> two possibilities :
>
> 1. We revert the JasperLoader changes back to parent.loadClass(name) call.
>
> 2. We modify Catalina to set the contextClassLoader during an include or a
> forward call.
>
> I still wonder which is the best approach, but I must say I don't quite
> understand why we have to store the class loader in the Thread. It seems a
> tad ugly to do it that way...
2 is the only correct one, since if you forward to a servlet, it wouldn't
work otherwise.
The fix is already in the CVS (you missed the commit message apparently).
I have a test case which I didn't add because it requires modifications to
the tester.
Remy
