Unfortunately, no. It's all proprietary stuff for my company. Sorry.
I'll be glad to help you with any problems you run into coding it up,
but I can't post mine. I can, however, tell you that I based my
interceptor factory on the interceptor factory for creating logging
I order to keep the code generation simple, I would also recommend
putting as much of the actual interception logic in a standard service
or class as possible, using the generated interceptor merely as a
hook. It greatly simplifies debugging.
On 5/5/06, Henri Dupre <henri.dupre@gmai...> wrote:
> On 5/3/06, Mike Snare <mikesnare@gmai...> wrote:
> > Henri,
> > This same issue has been discussed before (search the archives for my
> > name and authentication), but I'll summarize what I've done and how it
> > works -- and how it doesn't work.
> > I've created an authentication interceptor that I attach to all the
> > major engines. It simply checks for an authenticated user and throws
> > a RedirectException to the login page if none is found. It also tells
> > the login page where to go after a login. I have an external
> > configuration file that can be used by a page designer to flag that
> > his/her page should NOT be authenticated. By default, all pages are.
> > The problem is timing. The authenticator gets called before the page
> > is set up, and thus a callback can't be created. (I, too, ran into
> > the PageCallback issue trying to overcome this). There is an
> > enhancement request (TAPESTRY-892) that Jessee is looking at to enable
> > us to force the setup of a page specifically for this reason (and
> > others I haven't thought of).
> > This approach isn't perfect, but depending on how your pages operate,
> > does a good deal of the work for you.
> This sounds quite good. I was just thinking that these features would be
> nice to have directly inside the tapestry framework since I bet that it
> would suit the needs of 90% of people. I hope to see something happening
> about your enhancement request because this is also where I'm stuck right
> now. Is there any chance to see your code and hivemind configuration?