On Thursday 15 December 2005 19:48, Andrus Adamchik wrote:
> Hi Marek,
>
> On Dec 14, 2005, at 6:00 PM, Marek Wawrzyczny wrote:
> > It seems to me that every unique DataContext in 3T opens a new
> > HessianConnection, thus establishing a new Hessian session.
>
> Almost. From the *server* perspective each DataContext is what you
> call a "hessian session". From the client perspective, "hessian
> sessions" (same as server DataContexts) can be either dedicated (one
> per client connection) or shared by multiple clients. Still Hessian
> "sessionId" string passed around is unique for each connection, even
> if multiple ids share the same server DataContext.
>
> > However, this Hessian session is not tied to (and it doesn't appear
> > possible to tie it to)
> > the physical HttpSession.
>
> I am not that pessimistic. It should be possible to tie hessian
> "sessionId" to the HttpSession. This will require some Hessian
> hacking though. From what I can tell we need to subclass
> com.caucho.hessian.client.HessianProxy to handle cookies on the
> client (not sure about SSL certificates?).
>
> I'd say this whole thing warrants a Jira issue so that we don't loose
> track. Care to open one? Let's work on it.
I'll open a Jira issue sometime this week. I want to discuss this with Marcin
first as he is the one doing most of the work at present.
We still haven't explored SSL, I might have some time to look into this later
this week.
The cookie idea was our second choice. Marcin already thinks that he can hack
both server and client code relatively easily. We'll post here if we manage
to do that.
-- - Marek Wawrzyczny Send instant messages to your online friends http://au.messenger.yahoo.com
This archive was generated by hypermail 2.0.0 : Tue Dec 20 2005 - 16:03:37 EST