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 62129 invoked from network); 17 Oct 2000 01:46:51 -0000 Received: from dnai-216-15-97-206.cust.dnai.com (HELO shiva.betaversion.org) (216.15.97.206) by locus.apache.org with SMTP; 17 Oct 2000 01:46:51 -0000 Received: from [192.168.1.106] (dhcp-6.betaversion.org [192.168.1.106]) by shiva.betaversion.org (8.9.3+Sun/8.9.3) with ESMTP id SAA04915 for ; Mon, 16 Oct 2000 18:51:22 -0700 (PDT) User-Agent: Microsoft-Outlook-Express-Macintosh-Edition/5.02.2022 Date: Mon, 16 Oct 2000 18:46:51 -0700 Subject: Re: Proposed Plan: Tomcat 4.0-m3 on Wednesday From: "Pier P. Fumagalli" To: Message-ID: In-Reply-To: <39EBA5E0.6631C0E@eng.sun.com> Mime-version: 1.0 Content-type: text/plain; charset="US-ASCII" Content-transfer-encoding: 7bit X-Spam-Rating: locus.apache.org 1.6.2 0/1000/N Craig R. McClanahan wrote: > > One feature that is planned for m3, but not yet available in the CVS > sources (Pier is working hard on it :-) is a new web connector to Apache > that understands what a web application is. This will dramatically ease > the headaches of configuring the Apache connector for all the things > that are already specified in the web.xml file. Update on this... The connector is _COMPLETELY_ new... New protocol (it's a phoenix resurrecting from the old AJPv21 ashes), and the whole module is completely new (not a line of code from my old mod_jserv or mod_jk). There's going to be an explaination on WHY this _HAD_ to be done (at least from my point of view). > NOTE 1: This connector code should be considered EXPERIMENTAL at this > time -- it is definitely not ready for prime time, having been tested > only on Solaris8/x86 to date. We're counting on some help from people > willing to try it out in other environments and provide bug reports. > Once it settles down, the ports to the other web servers will be added > as well. Uh... It is kinda tested also under MacOS/X, single processor, since that's my development environment nowdays. As of _right_now_ I'm trying to abstract the connection layer, so that it'll be easier to add load balancing and connection encryption, and fighting with a couple of memory leaks and null pointers... (segmentation fault... oh my god how I love JAVA!). Also, I'm very sorry for not keeping the works in CVS, but I believe that on a piece of code written from scratch, where files come and go, design gets reconstructed once every two days, CVS is not the best solution... Pier