Hi Marc,
The fix for RemoteMethodInvocation is committed, along with
several other changes that allow the security principal and
credential to propagate in bean-to-bean calls, whether within a
single VM, or multiple VMs on multiple machines.
I had to make one small repair to the security metadata stuff. The
wildcard symbol for method name * can be specified to apply to
only the home interface or only the remote interface, but the
metadata assumed that a * always applied to all methods.
>From visual inspection of the code, it seems that there is a second
small error which I didn't fix. Permissions for a particular method
may be included in more than one <method-permission> element.
It seems in BeanMetaData that once a match is found, no more
matching elements would be considered. If he has a chance,
Sebastien might want to look at this. Otherwise, I'll try to fit it in
later this week or next weekend. It looks like a quick fix.
I'm ready for volunteers who want to help out with security stuff. A
few people have expressed casual interest. Take a look at bugzilla
#171 and tell me what you'd be interested in doing. I'll give you a
hand at getting started.
Thanks,
-Dan
On 21 Aug 00, at 9:21, marc fleury wrote:
>
> > > Hm.. that sounds fixable.. right?
> >
> > LOL. Of course I fixed it. I was just looking for sympathy. :-)
>
> we love you man we love you... please commit so I can work from a fresh base
> :)))
>
> marc
>
> >
> > -Dan
> >
> > >
> > > > There. I feel better now that I've shared. :-)
> > >
> > > Good! :-) Have you fixed it or should I..?
> > >
> > > /Rickard
> > >
> > > --
> > > Rickard �berg
> > >
> > > Email: [EMAIL PROTECTED]
> > > http://www.telkel.com
> > > http://www.jboss.org
> > > http://www.dreambean.com
> > >
> >
> >
> >
> >
>
>
