Return-Path: Delivered-To: apmail-tuscany-dev-archive@www.apache.org Received: (qmail 56971 invoked from network); 2 Aug 2009 21:38:20 -0000 Received: from hermes.apache.org (HELO mail.apache.org) (140.211.11.3) by minotaur.apache.org with SMTP; 2 Aug 2009 21:38:20 -0000 Received: (qmail 80321 invoked by uid 500); 2 Aug 2009 21:38:25 -0000 Delivered-To: apmail-tuscany-dev-archive@tuscany.apache.org Received: (qmail 80242 invoked by uid 500); 2 Aug 2009 21:38:24 -0000 Mailing-List: contact dev-help@tuscany.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: dev@tuscany.apache.org Delivered-To: mailing list dev@tuscany.apache.org Received: (qmail 80234 invoked by uid 99); 2 Aug 2009 21:38:24 -0000 Received: from nike.apache.org (HELO nike.apache.org) (192.87.106.230) by apache.org (qpsmtpd/0.29) with ESMTP; Sun, 02 Aug 2009 21:38:24 +0000 X-ASF-Spam-Status: No, hits=1.2 required=10.0 tests=SPF_HELO_PASS,SPF_NEUTRAL X-Spam-Check-By: apache.org Received-SPF: neutral (nike.apache.org: local policy) Received: from [212.227.126.177] (HELO moutng.kundenserver.de) (212.227.126.177) by apache.org (qpsmtpd/0.29) with ESMTP; Sun, 02 Aug 2009 21:38:14 +0000 Received: from [192.168.0.17] (93-97-220-25.zone5.bethere.co.uk [93.97.220.25]) by mrelayeu.kundenserver.de (node=mrelayeu4) with ESMTP (Nemesis) id 0ML21M-1MXika3s68-00018I; Sun, 02 Aug 2009 23:37:53 +0200 Message-ID: <4A760730.7030305@apache.org> Date: Sun, 02 Aug 2009 22:37:52 +0100 From: Simon Nash User-Agent: Thunderbird 2.0.0.22 (Windows/20090605) MIME-Version: 1.0 To: dev@tuscany.apache.org Subject: Re: [1.5.1] Domain manager and binding.sca References: <4A756138.2030500@apache.org> In-Reply-To: <4A756138.2030500@apache.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Provags-ID: V01U2FsdGVkX18iOfqkFmSFb6Hec2YCoxhAsW9rJJt2DujGBtS joZ2tBXoVEYGYcIkp2rMFZ7Zc6pqlUSJm4FYCExK5vCiMaKSuH bie0EczHg4k6PnZEhbP0g== X-Virus-Checked: Checked by ClamAV on apache.org Simon Nash wrote: > When creating a domain configuration from scratch using the > Domain Manager GUI, the generated configuration files don't include > a entry. This means that deployed composites that > use binding.sca for inter-node communication don't work "out of the > box" in a distributed domain configuration. In order to get this > to work, it's necessary to manually edit the generated domain > configuration to add elements to the node definitions. > > IMO this is a serious usability issue for someone just starting > with the distributed domain and I would like to change the domain > manager to generate a entry in the node configuration. > This is a simple one-line change and it doesn't cause any problems > in a full build. > > Adding this entry makes the distributed domain work "out of the box" > with binding.sca. However, with the current domain mananger code, > this change also removes the local optimization for inter-node calls > using . This is because the node configuration algorithm > is currently "all or nothing" which means that all references and > services using would get an http:// absolute URI, forcing > them to use binding-sca-axis2 for communication rather than binding-sca. > > The correct solution is to make the domain manager URI generation > algorithm more intelligent and only create the http:// style of > URI for binding.sca references and services that use inter-node wires, > not those that use intra-node wires. Unfortunately this change is > not simple because URI generation in the domain manager currently > precedes domain wiring, which means the URI generation code doesn't > know which references and service are wired together. > > I can see 3 options for 1.5.1: > 1. Don't add to node configurations in the domain > manager, and force users to do this manually. > 2. Add to node configurations in the domain manager, > and accept that there won't be any local optimization when > binding.sca is used in a distributed domain configuration. > 3. Add to node configurations in the domain manager, > and also fix the URI generation algorithm to preserve local > opimizations where they are applicable. I am currently thinking > about how best to implement this change. > > I am willing to work on option 3, but including this will delay the > release a little. I don't think option 1 is acceptable. I could live > with option 2, but others might not be happy with this. > > Comments, questions, opinions, preferences? > > Simon > > I have an idea for how to make option 3 work with only minor changes in the builders. I have written some code and I will test it tomorrow. Simon