Return-Path: Delivered-To: apmail-geronimo-dev-archive@www.apache.org Received: (qmail 31218 invoked from network); 23 Aug 2005 04:24:08 -0000 Received: from hermes.apache.org (HELO mail.apache.org) (209.237.227.199) by minotaur.apache.org with SMTP; 23 Aug 2005 04:24:08 -0000 Received: (qmail 90779 invoked by uid 500); 23 Aug 2005 04:24:01 -0000 Delivered-To: apmail-geronimo-dev-archive@geronimo.apache.org Received: (qmail 90737 invoked by uid 500); 23 Aug 2005 04:24:01 -0000 Mailing-List: contact dev-help@geronimo.apache.org; run by ezmlm Precedence: bulk list-help: list-unsubscribe: List-Post: Reply-To: dev@geronimo.apache.org List-Id: Delivered-To: mailing list dev@geronimo.apache.org Received: (qmail 90724 invoked by uid 99); 23 Aug 2005 04:24:01 -0000 Received: from asf.osuosl.org (HELO asf.osuosl.org) (140.211.166.49) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 22 Aug 2005 21:24:01 -0700 X-ASF-Spam-Status: No, hits=0.0 required=10.0 tests= X-Spam-Check-By: apache.org Received-SPF: neutral (asf.osuosl.org: local policy) Received: from [209.233.18.245] (HELO buttons.boynes.com) (209.233.18.245) by apache.org (qpsmtpd/0.29) with ESMTP; Mon, 22 Aug 2005 21:24:18 -0700 Received: from [192.168.37.199] (unknown [192.168.37.199]) by buttons.boynes.com (Postfix) with ESMTP id ED0F9194DF for ; Mon, 22 Aug 2005 21:23:57 -0700 (PDT) Message-ID: <430AA4DD.3090002@apache.org> Date: Mon, 22 Aug 2005 21:23:57 -0700 From: Jeremy Boynes User-Agent: Mozilla Thunderbird 1.0.6-1.1.fc3 (X11/20050720) X-Accept-Language: en-us, en MIME-Version: 1.0 To: dev@geronimo.apache.org Subject: Re: Web container-specific configuration References: <200508230051.j7N0pSPM016164@sun.savoirtech.com> <50a15c12ffb704af8681e3b7ef1cff66@yahoo.com> In-Reply-To: Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Virus-Checked: Checked by ClamAV on apache.org X-Spam-Rating: minotaur.apache.org 1.6.2 0/1000/N Aaron Mulder wrote: > I really disagree with having separate namespaces for the entire > web deployment plan for Tomcat and Jetty. It makes Geronimo+Tomcat and > Geronimo+Jetty totally different products. If I'm going to release a > typical application for Geronimo, you're saying that every single bit of > will be identical except for some stupid plumbing in the web plans? So > you must release a Geronimo+Tomcat version of the application and a > Geronimo+Jetty version of the application? Say it ain't so! > > I'll grant that it's possible to construct an application that > works properly in only one container or the other. But I really object to > crafting our whole configuration strategy around that case, which I expect > to be very rare. I think it's going to be much more common that a plan is > totally portable, or totally portable with a couple of container-specific > tweaks for both containers that don't cause the app to fail if not > deployed in its preferred container. I'd rather make that the baseline, > and allow a generic plan and a generic plan with extensions for 0-N web > containers. > I would assert that most applications will be targeted at one environment or another. Most enterprise users will pick one platform and develop for that, most ISVs will pick one platform and distribute for that. In either case the overhead of testing all possible combinations will prove too expensive. Let's face reality: Geronimo+Tomcat and Geronimo+Jetty *are* different products - they even require separate certification. Anyone distributing an application is either going to ship the one they work with, or ship pre-tested deployment plans for both (just like they ship plans for JBoss, WebLogic or WebSphere). The choice of web container is actually non-representative of the general case as there is relatively little web-container specific configuration. When you look at EJB configuration etc. there is much more container specific information needed - are we going to define one plan for all EJB containers? Say it ain't so! This experiment with common web container configuration has shown that implementations *are* different, even down to basic concepts like what a "host" is. And that's just with two sample points, what about others? -- Jeremy