Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
Unfortunately, we keep changing the artifact names and Plugin architecture/schema in every major release, so users who have developed real JavaEE apps on 2.0 (more than just simple servlet/jsp web apps) will more than likely have some work and testing before they can move to 2.1. If we want to build a vibrant user community, we need to support the prior major release (2.0) until our next major release (2.2) comes out. -Donald Jacek Laskowski wrote: On Thu, Feb 21, 2008 at 10:36 AM, Erik B. Craig [EMAIL PROTECTED] wrote: I have had this thought myself. Often times it seems we are far more focused on the next major release (I.E. current trunk) as opposed to introducing some more necessary fixes into other smaller releases. I think we should definitely continue to support / release on the 2.0 branch as much as possible, and would be happy to assist in integration of patches and changes that have been made since then. I always thought that our support for past releases has been the latest release. It will lower the burden greatly for us and for our users too. If we put more attention to documentation *and* release often with small changes they would benefit more. Jacek smime.p7s Description: S/MIME Cryptographic Signature
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
On Fri, Feb 22, 2008 at 9:21 AM, Donald Woods [EMAIL PROTECTED] wrote: If we want to build a vibrant user community, we need to support the prior major release (2.0) until our next major release (2.2) comes out. A vibrant community has nothing to do with the versions we support. If we provided enough information on the respective versions they'd love us more than when we support more versions than necessary. Users always have to verify whether their apps work or not irrespective of what exactly number in the version scheme has changed. I think they'd love if they could work with the latest and greatest stuff if they were ensured it would work with their apps. That's the real pain in any migration - lack of knowledge and experience. That's why Hernan and Co's work's so important. The more users know about Geronimo, the better for their migration efforts. Anyway, I'm only a lurker and if I fever find some spare cycles to work on Geronimo beside commenting out/responding to user questions I won't likely be fixing bugs/issues reported against the previous versions or branches, but rather enhance documentation/fix them in the trunk so our community is better equipped with the right tools for their job around Geronimo and to let them stay current with the changes. I keep saying it's always better for them and us. Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
Aka the 2.0 branch - https://svn.apache.org/repos/asf/geronimo/server/branches/2.0 We need to continue supporting our user community and stop forcing them to the latest and greatest release. Look at how Tomcat and the HTTP Server handle support with tons of . releases. I'm not saying we should have as many maintenance releases as Tomcat (ie. 5.5.26) but we should try to keep supporting a release by integrating patches for say 6 months after the last maintenance release, which would mean we're still in the 2.0.x 6 months window (released 20071019.) If a committer is willing to help solve a user's problem on a prior release, then +1000 to them. There are still users asking about 1.1.1 on the users mailing list, so obviously we need to do a better job of supporting our releases and stop this try the latest release and try the latest snapshot approach that we have been doing. My 2 cents -Donald Jacek Laskowski wrote: On Tue, Feb 19, 2008 at 3:37 PM, Vamsavardhana Reddy [EMAIL PROTECTED] wrote: With Geronimo Tomcat 2.0.3-SNAPSHOT, when a web app is stopped, I am noticing that the call to stop() in GeronimoStandardContext.kill() is making the sessions to be written to a SESSIONS.ser under the workDir for the application. But then destroy() called immediately is resulting in the deletion of the workDir altogether. Under what situations will this workDir be not deleted and how this SESSIONS.ser will be used/supposed to be used? What's 2.0.3-SNAPSHOT? I have never seen it mentioned before. Should we care about it rather than pushing 2.1 to our end users? Why are you working with the older version? Jacek smime.p7s Description: S/MIME Cryptographic Signature
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
BTW, this problem (or situation) applies to all versions of Geronimo. I have verified it on 1.1.1 and 2.0.3-SNAPSHOT versions. ++Vamsi On Thu, Feb 21, 2008 at 12:59 PM, Vamsavardhana Reddy [EMAIL PROTECTED] wrote: On Thu, Feb 21, 2008 at 11:14 AM, Jacek Laskowski [EMAIL PROTECTED] wrote: On Tue, Feb 19, 2008 at 3:37 PM, Vamsavardhana Reddy [EMAIL PROTECTED] wrote: With Geronimo Tomcat 2.0.3-SNAPSHOT, when a web app is stopped, I am noticing that the call to stop() in GeronimoStandardContext.kill() is making the sessions to be written to a SESSIONS.ser under the workDir for the application. But then destroy() called immediately is resulting in the deletion of the workDir altogether. Under what situations will this workDir be not deleted and how this SESSIONS.ser will be used/supposed to be used? What's 2.0.3-SNAPSHOT? I have never seen it mentioned before. It is the server built from branches\2.0. Should we care about it rather than pushing 2.1 to our end users? I think we should care about it too, for the users who are already using a 2.0.x server may not want to move to 2.1 until 2.1 stabilizes a bit. Why are you working with the older version? My Win XP crashes when I build geronimo-system module from 2.1. I could not figure the reason, but it has something to do with the model of my ThinkPad :o(. Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
On Thu, Feb 21, 2008 at 7:32 AM, Donald Woods [EMAIL PROTECTED] wrote: We need to continue supporting our user community and stop forcing them to the latest and greatest release. Look at how Tomcat and the HTTP Server handle support with tons of . releases. I'm not saying we should have as many maintenance releases as Tomcat (ie. 5.5.26) but we should try to keep supporting a release by integrating patches for say 6 months after the last maintenance release, which would mean we're still in the 2.0.x 6 months window (released 20071019.) ... There are still users asking about 1.1.1 on the users mailing list, so obviously we need to do a better job of supporting our releases and stop this try the latest release and try the latest snapshot approach that we have been doing. Well, it's an open source project and *I* don't have much time supporting the latest and gratest release not to mention I won't certainly have some for the past releases. The aim of our releases should always be to move forward without forgetting our past (and hopefully support people who use the past releases). Think of it this way - if we release 2.0.3, what would be a difference if we named it 2.0.10 or 2.1.1? If a user needs to fix an issue and has to upgrade to some version it doesn't really matter if it's 2.0.3, 2.0.58 or 2.1.1, right? That's what I'm talking about. Let's convince our users (or better yet let them know/believe) that moving with the latest and greatest release is for their better and calmer life. If an end user has to move up (s)he will have to test it out before upgrade, right? Does it really matter what version it will be if the latest one works? I don't think so. I keep hearing from my customers they don't like the latest release because it's not well baked and noone really knows what to expect. I keep answering them, well perhaps it is not but believe me your pain when something bad happens (after your app worked fine during testing) will be lower than when you're stuck with the past version which you can't fix or expect to be fixed soon. Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
On Thu, Feb 21, 2008 at 10:36 AM, Erik B. Craig [EMAIL PROTECTED] wrote: I have had this thought myself. Often times it seems we are far more focused on the next major release (I.E. current trunk) as opposed to introducing some more necessary fixes into other smaller releases. I think we should definitely continue to support / release on the 2.0 branch as much as possible, and would be happy to assist in integration of patches and changes that have been made since then. I always thought that our support for past releases has been the latest release. It will lower the burden greatly for us and for our users too. If we put more attention to documentation *and* release often with small changes they would benefit more. Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
Donald, I have had this thought myself. Often times it seems we are far more focused on the next major release (I.E. current trunk) as opposed to introducing some more necessary fixes into other smaller releases. I think we should definitely continue to support / release on the 2.0 branch as much as possible, and would be happy to assist in integration of patches and changes that have been made since then. Thanks, Erik B. Craig [EMAIL PROTECTED] On Feb 21, 2008, at 9:32 AM, Donald Woods wrote: Aka the 2.0 branch - https://svn.apache.org/repos/asf/geronimo/server/branches/2.0 We need to continue supporting our user community and stop forcing them to the latest and greatest release. Look at how Tomcat and the HTTP Server handle support with tons of . releases. I'm not saying we should have as many maintenance releases as Tomcat (ie. 5.5.26) but we should try to keep supporting a release by integrating patches for say 6 months after the last maintenance release, which would mean we're still in the 2.0.x 6 months window (released 20071019.) If a committer is willing to help solve a user's problem on a prior release, then +1000 to them. There are still users asking about 1.1.1 on the users mailing list, so obviously we need to do a better job of supporting our releases and stop this try the latest release and try the latest snapshot approach that we have been doing. My 2 cents -Donald Jacek Laskowski wrote: On Tue, Feb 19, 2008 at 3:37 PM, Vamsavardhana Reddy [EMAIL PROTECTED] wrote: With Geronimo Tomcat 2.0.3-SNAPSHOT, when a web app is stopped, I am noticing that the call to stop() in GeronimoStandardContext.kill() is making the sessions to be written to a SESSIONS.ser under the workDir for the application. But then destroy() called immediately is resulting in the deletion of the workDir altogether. Under what situations will this workDir be not deleted and how this SESSIONS.ser will be used/supposed to be used? What's 2.0.3-SNAPSHOT? I have never seen it mentioned before. Should we care about it rather than pushing 2.1 to our end users? Why are you working with the older version? Jacek
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
On Fri, Feb 22, 2008 at 12:12 AM, Jacek Laskowski [EMAIL PROTECTED] wrote: On Thu, Feb 21, 2008 at 7:32 AM, Donald Woods [EMAIL PROTECTED] wrote: We need to continue supporting our user community and stop forcing them to the latest and greatest release. Look at how Tomcat and the HTTP Server handle support with tons of . releases. I'm not saying we should have as many maintenance releases as Tomcat (ie. 5.5.26) but we should try to keep supporting a release by integrating patches for say 6 months after the last maintenance release, which would mean we're still in the 2.0.x 6 months window (released 20071019.) ... There are still users asking about 1.1.1 on the users mailing list, so obviously we need to do a better job of supporting our releases and stop this try the latest release and try the latest snapshot approach that we have been doing. Well, it's an open source project and *I* don't have much time supporting the latest and gratest release not to mention I won't certainly have some for the past releases. Of course it is open source and nobody is demanding anything from anyone :o). The aim of our releases should always be to move forward without forgetting our past (and hopefully support people who use the past releases). Think of it this way - if we release 2.0.3, what would be a difference if we named it 2.0.10 or 2.1.1? If a user needs to fix an issue and has to upgrade to some version it doesn't really matter if it's 2.0.3, 2.0.58 or 2.1.1, right? Migrating between major revisions is definitely more complex than migrating between minor revisions. With minor releases, it could be as simple as replacing a jar from the new release and making a few configuration changes to get the ball rolling. We may not add new features in minor releases, but we should definitely provide critical bug fixes. That's what I'm talking about. Let's convince our users (or better yet let them know/believe) that moving with the latest and greatest release is for their better and calmer life. If an end user has to move up (s)he will have to test it out before upgrade, right? Does it really matter what version it will be if the latest one works? I don't think so. I keep hearing from my customers they don't like the latest release because it's not well baked and noone really knows what to expect. I keep answering them, well perhaps it is not but believe me your pain when something bad happens (after your app worked fine during testing) will be lower than when you're stuck with the past version which you can't fix or expect to be fixed soon. The discussion on this thread has turned into something that is totally irrelevant to my original question. May be we should continue this discussion on a new thread?? Someone, who can really give me some inputs on what is happening with this SESSIONS.ser, please respond. Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Re: Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
On Tue, Feb 19, 2008 at 3:37 PM, Vamsavardhana Reddy [EMAIL PROTECTED] wrote: With Geronimo Tomcat 2.0.3-SNAPSHOT, when a web app is stopped, I am noticing that the call to stop() in GeronimoStandardContext.kill() is making the sessions to be written to a SESSIONS.ser under the workDir for the application. But then destroy() called immediately is resulting in the deletion of the workDir altogether. Under what situations will this workDir be not deleted and how this SESSIONS.ser will be used/supposed to be used? What's 2.0.3-SNAPSHOT? I have never seen it mentioned before. Should we care about it rather than pushing 2.1 to our end users? Why are you working with the older version? Jacek -- Jacek Laskowski http://www.JacekLaskowski.pl
Geronimo Tomcat 2.0.3-SNAPSHOT: SESSIONS.ser written to workDir
With Geronimo Tomcat 2.0.3-SNAPSHOT, when a web app is stopped, I am noticing that the call to stop() in GeronimoStandardContext.kill() is making the sessions to be written to a SESSIONS.ser under the workDir for the application. But then destroy() called immediately is resulting in the deletion of the workDir altogether. Under what situations will this workDir be not deleted and how this SESSIONS.ser will be used/supposed to be used? ++Vamsi