Return-Path: Mailing-List: contact tomcat-dev-help@jakarta.apache.org; run by ezmlm Delivered-To: mailing list tomcat-dev@jakarta.apache.org Received: (qmail 15104 invoked from network); 6 Jan 2000 09:35:54 -0000 Received: from typhoon.mail.pipex.net (158.43.128.27) by 63.211.145.10 with SMTP; 6 Jan 2000 09:35:54 -0000 Received: (qmail 15684 invoked from network); 6 Jan 2000 09:34:36 -0000 Received: from userar64.uk.uudial.com (HELO picasso.o3.co.uk) (62.188.136.223) by smtp.dial.pipex.com with SMTP; 6 Jan 2000 09:34:36 -0000 Received: by picasso.o3.co.uk (Postfix, from userid 0) id BC358D874; Thu, 6 Jan 2000 09:42:07 +0000 (GMT) Sender: mikew@picasso.o3.co.uk To: tomcat-dev@jakarta.apache.org Subject: Re: [LONG TERM PLAN] Proposed Architecture for Tomcat.Next Servlet Container References: <3866D000.2D544552@mytownnet.com> <3873A1A6.2685D1E2@eng.sun.com> <387416A9.6D01F5E4@mytownnet.com> From: Mike Williams Date: 06 Jan 2000 09:42:05 +0000 In-Reply-To: "Craig R. McClanahan"'s message of "Wed, 05 Jan 2000 20:14:34 -0800" Message-ID: Lines: 47 User-Agent: Gnus/5.0802 (Gnus v5.8.2) XEmacs/20.4 (Emerald) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii >>> "Costin" == Costin> Instead of a major version that will have all the features and Costin> changes you mentioned, I would prefer a number of minor changes Costin> that will make incremental changes. >>> "Craig" == "Craig R. McClanahan" Craig> I do not suggest that we build the Tomcat.Next interfaces on top Craig> of *any* existing Tomcat code. Instead, I advocate that we Craig> refactor and rebuild it. The existing code base is functional, Craig> but it has a lot of kludgey cruft in it, and almost no comments Craig> ... Craig> What do all the other Tomcat interested parties think? Firstly, respect to the JServ team for giving us a clean, stable Servlet platform. And respect to Sun (and the instigators within) for putting Tomcat under the Apache umbrella ... ensuring the success of Java Servlets. I'm a happy user of Apache JServ. Like Craig, I was frustrated by the amount of time it took for the Tomcat code to be released. My main complaint is that the Jakarta announcement very effectively stalled development of JServ ... if Jakarta hadn't happened, or the timeframe had be clearer, I'm guessing we'd now have a version of JServ that implemented version 2.2 (or at least 2.1) of the Servlet spec. Perhaps I'm wrong. My main interest is seeing a stable, complete, open-source implementation of Servlets-2.2, suitable for both development and production purposes. It seems that Tomcat is *almost* there (minus class re-loading), so it seems Reasonable to concentrate on that goal before attempting any major re-design. But once we're there, perhaps it's time to "re-factor" as Craig suggests, before adding additional functionality. If the existing Tomcat code is badly commented, it's going to be difficult to maintain, and messy to extend. Another advantage of "re-factoring" is that code can be lifted both from the current Tomcat and from JServ-1.1. Thus Tomcat-4.x would be an *evolution* of JServ, rather than just the recommended alternative.