cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Martin Man <>
Subject Re: Restricting objectModel
Date Fri, 18 May 2001 19:21:16 GMT
On Fri, May 18, 2001 at 08:20:29AM +0200, giacomo wrote:
> Dead developpers
> A recent question on the user list "how can I put my own stuff into the
> objectModel" forced me to rethink about the the concept of the
> objectModel.
> Let me resume: The Cocoon core engine doesn't really care about the
> environment it runs in. Thus we introduced the notion of an abstract
> Environment with the absolute minimum of functionality needed
> exclusively by the Cocoon engine to be able to work. To have a contract
> between a concrete Environment and the sitemap components, which are BTW
> the only component which need to know about the conrete Environment they
> run in, we introduced the objectModel as a container for important
> environmental objects. The concrete servlet Environment implemented
> today uses the objectModel to put in the Request, Response and Context
> objects of the servlet environment. This enables the Cocoon engine to
> operate in almost any environment (ie. Apache James [SMTP] or an EJB
> container) by simply creating an concrete Environment class.
> Now, I'd like to change the type of the objectModel from a Map to an
> Avalon Context. This will downgrade it to an read-only object preventing
> users from putting their own stuff in it.
+1 for it, when I was designing the validator actions I found out that putting
anything to objectModel actually does not help anything, ..., mainly because
flow is now controlled in a sitemap, so what matters is a Map returned to
the sitemap, because it can be then used from sitemap

the direct communication among components I consider obscure, because
1) it is not transparent and visible, you don't know what and who put there
and under which name
2) the state of the webapp is normally maintained using session, and we have now
session abstracted from the servlet environment, so writable object model is
just duplication and confusion here...

though, I might be wrong, 

> If nobody gives me a reasonable -1 I'd volunteer to change it the way
> I've proposed.
> Please your quick comments about it.
> Giacomo
"Only dead fish swims with a stream"
gpg_key_fingerprint: 2CC0 4AF6 92DA 5CBF 5F09  7BCB 6202 7024 6E06 0223

To unsubscribe, e-mail:
For additional commands, email:

View raw message