Return-Path: Delivered-To: apmail-ws-axis-dev-archive@www.apache.org Received: (qmail 31548 invoked from network); 10 Oct 2007 21:42:09 -0000 Received: from hermes.apache.org (HELO mail.apache.org) (140.211.11.2) by minotaur.apache.org with SMTP; 10 Oct 2007 21:42:09 -0000 Received: (qmail 18927 invoked by uid 500); 10 Oct 2007 21:41:50 -0000 Delivered-To: apmail-ws-axis-dev-archive@ws.apache.org Received: (qmail 18862 invoked by uid 500); 10 Oct 2007 21:41:50 -0000 Mailing-List: contact axis-dev-help@ws.apache.org; run by ezmlm Precedence: bulk Reply-To: axis-dev@ws.apache.org list-help: list-unsubscribe: List-Post: List-Id: Delivered-To: mailing list axis-dev@ws.apache.org Received: (qmail 18851 invoked by uid 99); 10 Oct 2007 21:41:50 -0000 Received: from athena.apache.org (HELO athena.apache.org) (140.211.11.136) by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 10 Oct 2007 14:41:50 -0700 X-ASF-Spam-Status: No, hits=-2.8 required=10.0 tests=RCVD_IN_DNSWL_MED,SPF_NEUTRAL X-Spam-Check-By: apache.org Received-SPF: neutral (athena.apache.org: local policy) Received: from [64.18.13.41] (HELO chip2og52.obsmtp.com) (64.18.13.41) by apache.org (qpsmtpd/0.29) with ESMTP; Wed, 10 Oct 2007 21:41:52 +0000 Received: from source ([192.150.11.134]) by chip2ob52.postini.com ([64.18.5.12]) with SMTP; Wed, 10 Oct 2007 14:39:05 PDT Received: from inner-relay-1.corp.adobe.com ([153.32.1.51]) by outbound-smtp-1.corp.adobe.com (8.12.10/8.12.10) with ESMTP id l9ALbEIQ016601 for ; Wed, 10 Oct 2007 14:37:14 -0700 (PDT) Received: from fe1.corp.adobe.com (fe1.corp.adobe.com [10.8.192.70]) by inner-relay-1.corp.adobe.com (8.12.10/8.12.10) with ESMTP id l9ALd4RC016696 for ; Wed, 10 Oct 2007 14:39:04 -0700 (PDT) Received: from namail4.corp.adobe.com ([10.8.192.68]) by fe1.corp.adobe.com with Microsoft SMTPSVC(6.0.3790.1830); Wed, 10 Oct 2007 14:39:05 -0700 x-mimeole: Produced By Microsoft Exchange V6.5 Content-class: urn:content-classes:message MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Subject: RE: Getting axis2 transport out from the kernel Date: Wed, 10 Oct 2007 14:39:03 -0700 Message-ID: In-Reply-To: X-MS-Has-Attach: X-MS-TNEF-Correlator: Thread-Topic: Getting axis2 transport out from the kernel Thread-Index: AcgKlyvmi/OxTiD+QGabfedrwcnvBgA7WNEA References: <4705E8DC.3020506@wso2.com> <470AEF19.8060405@opensource.lk> From: "Tom Jordahl" To: X-OriginalArrivalTime: 10 Oct 2007 21:39:05.0004 (UTC) FILETIME=[FB053EC0:01C80B85] X-Virus-Checked: Checked by ClamAV on apache.org > One jar containing the http transport (which can be > included in the axis2 standard bin) > And another jar containing the rest of the transports. +1 to moving out the other transports (everything but HTTP) from = axis2.jar to optional-transports.jar. I would break my 'too many jars = are bad' rule for things like that. -1 (disagreement, not formal veto) to removing the HTTP transport from = axis2.jar. It is configurable in the XML, right? Why can't Synapse = just configure their axis to use their transport? Why would we want = anyone to have org.apache.axis2.HTTPTransport configured that *wasn't* = the actual Axis2 HTTP transport, but something that was = better/cooler/faster/had different bugs? That could get confusing real = fast. I may be way off base with exactly what the use case to replacing the = 'standard' http transport in Axis2 is, if so then forgive my denseness. = :-)=20 -- Tom Jordahl ________________________________________ From: Rajith Attapattu [mailto:rajith77@gmail.com]=20 Sent: Tuesday, October 09, 2007 1:07 PM To: axis-dev@ws.apache.org Subject: Re: Getting axis2 transport out from the kernel Tom has a point. TCP and SMTP seems to be used very rarely. Perhapse < = 1%. JMS seems to be more popular.=20 But I still think we should to put transports in a separate module. When we make the axis2 standard distribution can we not just simply make = two jars out of the transport module?=20 One jar containing the http transport (which can be included in the = axis2 standard bin) And another jar containing the rest of the transports. Is this possible with maven? This way we can allow people to download the additional transports only = if they want to.=20 Synapse folks are happy bcos they can easily import only the axis2 = kernal and bundle their own HTTP transport. Does this sound reasonable? Regards, Rajith Attapattu Red Hat. On 10/9/07, Tom Jordahl wrote: > Well the best thing, IMO, is to ship the current http transport > with the release, by default and let users download others if > they want to use. +1 - The best thing for Axis2 is for the Axis users, not for Synapse.=20 The fewer jar files the better as well.=A0=A0The core of Axis2 should = have the 'standard' HTTP transport - anything else is just noise to 90+% of users. Keep your eyes on the target: making Axis2 easy to use for users.=20 -- Tom Jordahl -----Original Message----- From: Eran Chinthaka [mailto:chinthaka@opensource.lk] Sent: Monday, October 08, 2007 11:02 PM To: axis-dev@ws.apache.org Cc: synapse-dev@ws.apache.org Subject: Re: Getting axis2 transport out from the kernel -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1=20 Ok, I have couple of questions just to understand what the plan is. Let's assume we separate out transports in to a module. (begin What will this module contain? Will it have all the transports,=20 including http, smtp, tcp, etc., or will there be modules for each and every transport? What will be shipped with Axis2 release? All of them or none? If it is none, =A0=A0=A0=A0=A0=A0=A0=A0do you expect people to download Axis2 release = AND another=20 transport that he/she wants to use? else will we release the transport we have now? =A0=A0=A0=A0=A0=A0=A0=A0if yes, then Synapse will again have the same = problem. What will be inside default axis2.xml? Well the best thing, IMO, is to ship the current http transport with the = release, by default and let users download others if they want to use. The intuition is that most of the time the transport will be http and will make the life easier for about 90% of the users. Also I'd like to evaluate the impact of this on the HTTP binding=20 implementation that we have now. I think that also will go out of Kernel. ) Just my 2 cents. Thanks, Chinthaka Srinath Perera wrote: >> Guys.. the confusion/problem/root cause is this.. the Synapse=20 community >> re-wrote the JMS transport initially, and checked this into the Axis2 >> Kernel module as other transports were in there at the time. Actually >> this replaced the previous JMS transport Axis2 had.=20 >> >> Then we wrote a new non-blocking http/s transport, and kept this within >> Synapse. After some time there was a suggestion from Axis2 that we >> should check this code into Axis2 instead to make it available to a=20 >> wider audience for usage, testing and improving overall quality as well. >> Thus we checked this into Axis2 [Kernel :-(] > +1 yes that is right way to go > >> Now the problem is this.. The JMS transport was implemented for JMS=20 1.1, >> and the ESB community encountered users who wanted to deploy production >> apps using JMS 1.0.2. We also had a few bug fixes going into the NIO >> transport - where we couldn't wait for "the next axis2" release. We=20 also >> developed a Apache VFS based file/ftp/sftp/.. transport lately. We would >> like to give this code to all Axis2 users for sure, but not within the >> Axis2-kernel module - as it would create issues for us to make never=20 >> versions of these same classes for our use. >> >> As you may be already aware, Synapse is very close to transports - i.e. >> we want the maximum out of them, and would encounter most users who=20 >> would be looking for custom transports etc. We also do lots of load >> testing for concurrency and NIO related issues.. and we want to stay >> upto-date with the latest HttpCore release. Now since we committed=20 our >> initial transport code into Axis2-kernel, we are in a fix on making safe >> updates as and when required for Synapse releases. >> >> All we ask for is for Axis2 to create a module for transports and=20 take >> transports out from the kernel. > > > OK you convinced me. I am +1 on creating module for transports. > However as David pointed out we have to do it in Axis2 major release. > > One middle ground would be to create a transport module and first add > additional transports while leaving what ever we have in kernel as it > is for now. At a major release you can move other transports out as=20 > well. (May be you can leave http transport in the core for testing > reasons, but that we can decide later). That way we can ship > additional transports with minor release without adding them to > axis2.xml, and synapse can add them via custom axis2.xml. > > Thanks > Srinath > > >> asankha >> >> >> --------------------------------------------------------------------- = >> To unsubscribe, e-mail: axis-dev-unsubscribe@ws.apache.org >> For additional commands, e-mail: axis-dev-help@ws.apache.org=20 >> >> > > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org=20 iD8DBQFHCu8ZjON2uBzUhh8RAmbkAJ9fom4XImV2+x+7iPmBrF+rzC+l/ACcCckk Z3saNc4pBfUWn7OsV/pFzGo=3D =3Dwlxq -----END PGP SIGNATURE----- ---------------------------------------------------------------------=20 To unsubscribe, e-mail: axis-dev-unsubscribe@ws.apache.org For additional commands, e-mail: axis-dev-help@ws.apache.org=20 --------------------------------------------------------------------- To unsubscribe, e-mail: axis-dev-unsubscribe@ws.apache.org For additional commands, e-mail: axis-dev-help@ws.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: axis-dev-unsubscribe@ws.apache.org For additional commands, e-mail: axis-dev-help@ws.apache.org