cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carsten Ziegeler" <cziege...@sundn.de>
Subject AW: Restricting objectModel
Date Tue, 22 May 2001 08:49:29 GMT
> Giacomo Pati wrote:
>
> On Mon, 21 May 2001, Vadim Gritsenko wrote:
>
> > Carsten,
> >
> > I did try request attributes with <jsp:forward> - everything
> works fine for me;
> > I went through the weblogic.servlet.internal.ServletRequestImpl
> and RequestAttributes
> > (same package) - and, yes, there is check "instanceof
> Serializable", but no, all other
> > attributes are also alowed (just stored in another Map).
>
> I'm confused now. Is it possible for weblogic users to use
> request.setAttribute without the need to make the objects put there
> Serializable?
>
Yes, it is. But as our tests proved, there are problems if the
objects are not serializable. The test case for this is rather
complicated:
The servlet (cocoon) processes a request, we but some not Serializable
objects into the request attributes. Now the current thread starts
some (sub)threads and waits for them to finish.
And this is the cause for the problems. It seems (I am not an WLS expert),
that when the main threads starts sleeping the request object will be
serialized (don't ask me why). When the sub threads finish and the main
thread awakes the request object is deserialized. This serialization
is performed with all attributes of the request, regardless if they
are serializable or not (this is what the exception says).

So the usual test cases will work as the main thread never sleeps.

Carsten

> Giacomo
>
> >
> > PS My test is attached, run Test.jsp. If you can, please give
> me non-working test.
> >
> > Vadim
> >
> > > -----Original Message-----
> > > From: Carsten Ziegeler [mailto:cziegeler@sundn.de]
> > > Sent: Monday, May 21, 2001 1:41
> > > To: cocoon-dev@xml.apache.org
> > > Subject: AW: Restricting objectModel
> > >
> > >
> > > > Von: Vadim Gritsenko [mailto:vgritsenko@hns.com]
> > > >
> > > > Carsten,
> > > >
> > > > Do you use WL6.0 or WL6.0sp1?
> > > >
> > > WL6.0sp1
> > >
> > > Carsten
> > >
> > > > Vadim
> > > >
> > > > > -----Original Message-----
> > > > > From: Matthew Langham [mailto:mlangham@sundn.de]
> > > > > Sent: Friday, May 18, 2001 9:44
> > > > > To: cocoon-dev@xml.apache.org
> > > > > Subject: AW: Restricting objectModel
> > > > >
> > > > >
> > > > > Vadim,
> > > > >
> > > > > >>
> > > > > And about request attributes - that may be done for
> automatic failover /
> > > > > load
> > > > > balancing - I'm not so sure why is this done this way...
> > > > > <<
> > > > > Carsten was only talking about the request attributes - not the
> > > > session. The
> > > > > serialization of request attributes seems to be a "feature" of
> > > > Bea WLS 6.0
> > > > > as we have not seen it on any other application server we
> have installed
> > > > > onto.
> > > > >
> > > > > Matthew
> > > > >
> > > > > --
> > > > > Open Source Group               sunShine - Lighting up e:Business
> > > > > =================================================================
> > > > > Matthew Langham, S&N AG, Klingenderstrasse 5, D-33100 Paderborn
> > > > > Tel: +49-5251-1581-30   [mlangham@sundn.de - http://www.sundn.de]
> > > > > =================================================================
> > > > >
> > > > >
> > > > > -----Ursprungliche Nachricht-----
> > > > > Von: Vadim Gritsenko [mailto:vgritsenko@hns.com]
> > > > > Gesendet: Freitag, 18. Mai 2001 15:28
> > > > > An: cocoon-dev@xml.apache.org
> > > > > Betreff: RE: Restricting objectModel
> > > > >
> > > > >
> > > > > Carsten,
> > > > >
> > > > > </skip>
> > > > >
> > > > > > > > >
> > > > > > > > I forget some important point:
> > > > > > > > I know that the "correct" way of passing
> information should not be
> > > > > > > > the objectModel. You will tell me in your next email
that
> > > > I could use
> > > > > > > > the session or the request object for this, right?
> > > > > > >
> > > > > > > Absolutely :)
> > > > > > >
> > > > > > > > The problem with this is that some servlet engines
(like
> > > > BEA WLS 6.0)
> > > > > > > > requires that even the attributes of the request
> are serializable.
> > > > > > >
> > > > > > > Shit, is this a standard?
> > > > > >
> > > > > > Not sure about the standard, but BEA WLS 6.0 is known
> for exactly this
> > > > > > (there are discussions about this over on in the bea
> newsgroups.)
> > > > > > Our tests did prove this.
> > > > >
> > > > > You do not need to perform any tests to confirm this: If
> you think about
> > > > > running
> > > > > your application in clustered environment and remember that
> > > > serialization is
> > > > > the
> > > > > only way of replicating your session into backup server - than
> > > > it would be
> > > > > obvious
> > > > > to you why it's done this way :)
> > > > >
> > > > > And about request attributes - that may be done for
> automatic failover /
> > > > > load
> > > > > balancing - I'm not so sure why is this done this way...
> > > > >
> > > > > So, may be, you want your objects to be serializable to run
> > > > well in cluster.
> > > > >
> > > > > Regards,
> > > > > Vadim
> > > > >
> > > > >
> > > > > > >
> > > > > > > > And this prevents from putting anything into the
> request. Now we
> > > > > > > > could say: "Ok, that's a problem of the servlet engine
> > > > but not ours"
> > > > > > > > but this is of course no solution.
> > > > > > > > If there are really many votes with "+1" for
> changing to an avalon
> > > > > > > > context I would vote "+10000" for a container
> object which I can
> > > > > > > > get from the context to put any objects into for the
> > > > current request.
> > > > > > >
> > > > > > > Hmm..
> > > > > > >
> > > > > > >
> > > > > > > Can I read in between your lines that you already us
> it that way?
> > > > > >
> > > > > > Yes. And changing it would hit us hard.
> > > > > >
> > > > > >
> > > > > > Carsten
> > > > > >
> > > > > > Open Source Group                        sunShine - b:Integrated
> > > > > > ================================================================
> > > > > > Carsten Ziegeler, S&N AG, Klingenderstrasse 5, D-33100 Paderborn
> > > > > > www.sundn.de                          mailto: cziegeler@sundn.de
> > > > > > ================================================================
> > > > >
> > > > >
> > > > >
> ---------------------------------------------------------------------
> > > > > To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> > > > > For additional commands, email: cocoon-dev-help@xml.apache.org
> > > > >
> > > > >
> > > > >
> ---------------------------------------------------------------------
> > > > > To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> > > > > For additional commands, email: cocoon-dev-help@xml.apache.org
> > > > >
> > > > >
> > > >
> > > >
> ---------------------------------------------------------------------
> > > > To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> > > > For additional commands, email: cocoon-dev-help@xml.apache.org
> > > >
> > >
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> > > For additional commands, email: cocoon-dev-help@xml.apache.org
> > >
> > >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
> For additional commands, email: cocoon-dev-help@xml.apache.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: cocoon-dev-unsubscribe@xml.apache.org
For additional commands, email: cocoon-dev-help@xml.apache.org


Mime
View raw message