|> |> 1- is java: == local a violation of the j2ee spec? (i.e RMH
|java: lookups
|> |> from the client wont' work right?
|> |
|> |No, see that's another story. The java: for clients will be local, *but*
|> |the bindings in the java: namespace on the client will point to a remote
|> |JNDI (through a LinkRef).
|>
|> pfffff
|>
|> you need to put the LinkRef you get from some place, please what is that
|> place...
|> or explain what you are thinking about... if an application
|deployer is to
|> set up the visibility of the java: namespaces then that needs to
|include the
|> client and I can't see how your "put a linkref" works for every VM on the
|> planet.
|>
|> is there anything that you read in spec that says java: *is* VM local (is
|> there? I don't know honestly, let me know)
|
|Not quite following. A bean will be bound into the (distributed!)
|default namespace. Let's say it is bound to "someapp/MyBean". In a
|client we want "java:comp/env/ejb/MyBean" to be bound to that, so we
|simply bind "java:comp/env/ejb/MyBean" in the client to a LinkRef with
|the ref "someapp/MyBean". When the client looks up
|"java:comp/env/ejb/MyBean" it will be dereferenced to the binding in the
|distributed namespace, i.e. someapp/MyBean, which will return the beans
|home.
|
|R U with me?
Listen I need to go read the code, but from what I understand yes that works
on the server, how that works on a client somewhere in the web (RMH) is
beyond me. I simply don't see that. You don't have the knowledge to do the
construction you just described on the *client*. You can't build your java:
namespace from the client, with LinkRef, period.
|
|> |> 3- be explicit then (thanks to the one that pointed out the correct
|> |> spelling, that will teach me to go loud :) so if you want to
|put all your
|> |> local stuff either put it in local: (like I want a system:) or do a
|> |> java:/comp/env/local/ whatever but name that "locality" if you must
|> |
|> |How about java:server/? E.g. java:server/TransactionManager and
|> |java:server/jdbc/MyPool
|>
|> that is good but again when we move to cluster the fact that java: is VM
|> dependent kills the whole idea...
|
|NO NO NO... we have *2* namespaces... sorry, I should have been more
|clear on that. We will have the default namespace, as we do today, for
|everything that needs to be accessible from outside, i.e. EJB homes. For
|VM local stuff we use java:. Internal/external, java:/default. (and when
|I say "default" I mean "what you get when you do new InitialContext()").
again java: on ejbHomes is missing...
|> |Wrong implementation!? What is wrong with the implementation?? From what
|> |I can deduce above your only doubt is the name, i.e. "java:". Otherwise
|> |it fits what we want.
|>
|> relax, I should have been more clear. I am not criticism the "code" just
|> that the idea of scoping VMs is smaller than server: for example and that
|> RMH clients won't work with that... the code is really good
|though :))) I
|> really like it :)))
|
|See above. It Will Work(tm). I think the confusion was because you
|thought I was removing the current distributed JNDI namespace.. which is
|not the case, 'cause then nothing would work.. :-)
Don't worry about it, my point is that java: might be a more powerful beast,
but we will get there when we get there.
hey I mean it you are truly advancing our beat now, so really don't worry
about it... there is so much road to cover.
marc
|
|/Rickard
|
|--
|Rickard �berg
|
|Email: [EMAIL PROTECTED]
|http://www.telkel.com
|http://www.jboss.org
|http://www.dreambean.com
|
|
|
