Return-Path: Delivered-To: apmail-incubator-geronimo-dev-archive@incubator.apache.org Received: (qmail 78615 invoked by uid 500); 9 Aug 2003 03:28:30 -0000 Mailing-List: contact geronimo-dev-help@incubator.apache.org; run by ezmlm Precedence: bulk list-help: list-unsubscribe: list-post: Reply-To: geronimo-dev@incubator.apache.org Delivered-To: mailing list geronimo-dev@incubator.apache.org Received: (qmail 78598 invoked from network); 9 Aug 2003 03:28:29 -0000 Received: from wwws.ppgia.pucpr.br (HELO hermes.ppgia.pucpr.br) (200.192.112.141) by daedalus.apache.org with SMTP; 9 Aug 2003 03:28:29 -0000 Received: from sheridan (hercules.ppgia.pucpr.br [200.192.112.140]) by hermes.ppgia.pucpr.br (Postfix) with ESMTP id 350502B0030 for ; Sat, 9 Aug 2003 00:40:11 -0300 (BRT) From: "Denes" To: Subject: RES: RES: Dynamic proxies Date: Sat, 9 Aug 2003 00:39:00 -0300 Message-ID: <000001c35e27$c57c2990$0301a8c0@sheridan> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook, Build 10.0.2627 In-Reply-To: <20030808210737.1DBE72B0066@hermes.ppgia.pucpr.br> X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2800.1165 Importance: Normal X-Spam-Rating: daedalus.apache.org 1.6.2 0/1000/N > -----Mensagem original----- > De: a.vasconcelos@att.net [mailto:a.vasconcelos@att.net] > Enviada em: sexta-feira, 8 de agosto de 2003 17:56 > Para: denes@ppgia.pucpr.br > Cc: geronimo-dev@incubator.apache.org > Assunto: Re: RES: Dynamic proxies >=20 > Ok ... I finaly now have both email addresses in the list :o) >=20 > One example at a time. >=20 > 1. Can you define "use" for me ? I think that making a remote call to > an > EJB deployed in a different EAR constitutes a "use" relationship right ? > In > that case number 1 is not a problem if you consider separate EARs as > different applications. Did I miss something ? :o\ Yes. This is a problem because I want to call the Local, and not the Remote Interface of the EJB that is in another EAR. >=20 > 2. What do you mean by "I tried to do the same thing in several jars > inside a single ear, but all applications servers that I tried stops and > restars all EJBs in the ear file when updating" ? If you are replacing the > EAR with a new one with the new JARs in it, the application server will > have > to stop the application right ? I am not sure what you meant by that > phrase. For example, I have a system that integrates several legacy applications (SAP, CICS and DL1, among others) and performs some processing on these data. You can see this application as a core module, which has the common features that almost all other modules uses, and the other modules that perform some specific tasks. By separating this way, I can update almost every module other than core without stopping the overall application. So, if I want to update the DL1 module, while it's restarting the CICS and SAP modules are still working. I can also add other modules to integrate the remaining legacy systems without stopping the application as a whole. Hope that's clear. My English in not enough to explain it better, sorry... >=20 > I still think that keeping the concept of an application and of modules > within the application as the spec has is the way to go. > Alex V. >=20 Can you tell me where in the specs is an ear file defined? I overlooked the 1.4 spec and EJB 2.0 spec and can=B4t find it. Thanks. Denes > -- > "Many pilots buy performance and > think they are buying skill." Ken > Stewart > > That's not exactly the point that I had in mind. I had to work with this > issue > > of multiple ears to create an application in several projects, basically > to > > solve 2 problems: > > > > 1. When you have an application that has to use EJBs that are already > created > > for another application in the same corporation. I prefer the idea that > this > > application simply uses the EJBs of another ear than have to redeploy > all the > > classes in each ear. > > > > 2. When you have a large application that has several modules that work > > togheter, several ears optimize the process of hot deployment, as the > > application server will replace only the EJBs of the ear being updated. > (I > > tried to do the same thing in several jars inside a single ear, but all > > applications servers that I tried stops and restars all EJBs in the ear > file > > when updating). > > > > Denes > > > > Citando a.vasconcelos@att.net: > > > > > Not sure if I agree with this one, although I think it is mainly an > > > academic issue. >=20 > > > The way I read the spec the EAR IS indeed what sets the boundary of > an > > > application. The different JAR and WAR files set the boundaries for > sub- > > > systems and possibly business objects. In other words, different > concerns of > > > > > > an overall application are bundled inside the EAR while interations > between > > > > > > EARs show dependencies between disparate applications that are not > > > necessarily part of an "overall application" ... unless of course we > want to > > > > > > end up with a grouping of every EAR ever written since one way or > another one > > > > > > could argue that they are all part of the same "overall" scheme :o) > > > > > > Sharing classes between EARs is IMHO counter-intuitive. > > > > > > Alex V. > > > > > > -- > > > "Many pilots buy performance and > > > think they are buying skill." Ken > > > Stewart > > > > IMHO this is not a so huge mistake... If you think about the J2EE > > > > architecture, one of the goals is to provide a set of business >=20 > > > > components, which can be (ideally) integrated to create an > application. > > > > > > > > When you have several ears in a server, one could argue that these > ears > > > > are different concerns about different features of the overall > > > > application. The application is not the ear, but the interaction of > the > > > > several components that are in each ear file. > > > > > > > > If you want to isolate components, you should create different > server > > > > configurations to isolate them. > > > > > > > > Denes > > > > > > > > > -----Mensagem original----- > > > > > De: Michael Remijan [mailto:remijan@ncsa.uiuc.edu] > > > > > Enviada em: sexta-feira, 8 de agosto de 2003 12:43 > > > > > Para: geronimo-dev@incubator.apache.org; > > > > geronimo-dev@incubator.apache.org > > > > > Assunto: Re: Dynamic proxies > > > > > > > > > > JBoss redid their classloader for version 3 and it is very > > > > > counter-intuitive. Basically, unless you use a jboss specific > config > > > > > file, > > > > > when you deploy an EAR it will share classes from other ears! > IMNSHO > > > > this >=20 > > > > > is a huge mistake. The JBoss documentation says the classloaders > were > > > > > redesigned this way to more fully incorporate JMX. > > > > > > > > > > > > > > > > > > -- > > Existem 10 tipos de pessoas: as que entendem bin=E1rio e as que = n=E3o. > > > > ------------------------------------------------- > > This mail sent through IMP: http://horde.org/imp/