Re: Maven Core to Git
2012/11/27 Brett Porter br...@apache.org: On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. +1. That will make that a sort of Maven tck. - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Olivier Lamy Talend: http://coders.talend.com http://twitter.com/olamy | http://linkedin.com/in/olamy - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
On 27 November 2012 08:41, Olivier Lamy ol...@apache.org wrote: 2012/11/27 Brett Porter br...@apache.org: On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. +1. That will make that a sort of Maven tck. Which is what it should be IMHO - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Olivier Lamy Talend: http://coders.talend.com http://twitter.com/olamy | http://linkedin.com/in/olamy - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
Didn't it used to be that way? (separate) On Tue, Nov 27, 2012 at 4:09 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On 27 November 2012 08:41, Olivier Lamy ol...@apache.org wrote: 2012/11/27 Brett Porter br...@apache.org: On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. +1. That will make that a sort of Maven tck. Which is what it should be IMHO - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Olivier Lamy Talend: http://coders.talend.com http://twitter.com/olamy | http://linkedin.com/in/olamy - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
Yes, but Kristian suggested merging core and its ITs and we all agreed initially but now I think it's more trouble than its worth. At least to start. On Nov 27, 2012, at 10:52 AM, Brian Fox bri...@infinity.nu wrote: Didn't it used to be that way? (separate) On Tue, Nov 27, 2012 at 4:09 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On 27 November 2012 08:41, Olivier Lamy ol...@apache.org wrote: 2012/11/27 Brett Porter br...@apache.org: On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. +1. That will make that a sort of Maven tck. Which is what it should be IMHO - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Olivier Lamy Talend: http://coders.talend.com http://twitter.com/olamy | http://linkedin.com/in/olamy - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - What matters is not ideas, but the people who have them. Good people can fix bad ideas, but good ideas can't save bad people. -- Paul Graham
Re: Maven Core to Git
https://issues.apache.org/jira/browse/INFRA-5390 has been updated to keep a separate it repo. The issue should be handled by infra real soon now. Kristian 2012/11/27 Jason van Zyl ja...@tesla.io: Yes, but Kristian suggested merging core and its ITs and we all agreed initially but now I think it's more trouble than its worth. At least to start. On Nov 27, 2012, at 10:52 AM, Brian Fox bri...@infinity.nu wrote: Didn't it used to be that way? (separate) On Tue, Nov 27, 2012 at 4:09 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On 27 November 2012 08:41, Olivier Lamy ol...@apache.org wrote: 2012/11/27 Brett Porter br...@apache.org: On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. +1. That will make that a sort of Maven tck. Which is what it should be IMHO - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Olivier Lamy Talend: http://coders.talend.com http://twitter.com/olamy | http://linkedin.com/in/olamy - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - What matters is not ideas, but the people who have them. Good people can fix bad ideas, but good ideas can't save bad people. -- Paul Graham - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Maven Core to Git
Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance
Re: Maven Core to Git
The problem is the asf builds running too often, and sometimes far too often and hence spamming the mailing lists substantially. Some of it has been solved, but there are still a few issues remaniing: As far as I can see there are two aspects of the issue: 1. A configuration error (or any kind of git corruption/inconsistency) on any of the git nodes would lead to the jenkins build running non-stop, generating mass amounts of spam on the notifications mailing lists. This issue has been solved as far as I can see; both by fixing the git configuration issue and by patching jenkins to fix the issue. This was https://issues.jenkins-ci.org/browse/JENKINS-15803. 2. There is a second issue where any kind of intermittent disconnect between the main jenkins instance and its node will trigger a rebuild because the master does not distinguish between a node being blank/reconfigured and simply unavaliable at the moment. This is basically happening by workspaceOffline returning true in https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L1437 I've been meaning to get Olivier to add logging in this code to find out what is actually happening. It seems to me like every single glitch in the asf network is causing rebuilds. And it would appear to be a quite unstable network Kristian 2012/11/26 Jason van Zyl ja...@tesla.io: Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
On Nov 26, 2012, at 12:47 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: The problem is the asf builds running too often, and sometimes far too often and hence spamming the mailing lists substantially. Some of it has been solved, but there are still a few issues remaniing: As far as I can see there are two aspects of the issue: 1. A configuration error (or any kind of git corruption/inconsistency) on any of the git nodes would lead to the jenkins build running non-stop, generating mass amounts of spam on the notifications mailing lists. This issue has been solved as far as I can see; both by fixing the git configuration issue and by patching jenkins to fix the issue. This was https://issues.jenkins-ci.org/browse/JENKINS-15803. 2. There is a second issue where any kind of intermittent disconnect between the main jenkins instance and its node will trigger a rebuild because the master does not distinguish between a node being blank/reconfigured and simply unavaliable at the moment. This is basically happening by workspaceOffline returning true in https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L1437 I've been meaning to get Olivier to add logging in this code to find out what is actually happening. It seems to me like every single glitch in the asf network is causing rebuilds. And it would appear to be a quite unstable network So in practical terms it all needs to be released, and the servers all updated? Can we just turn off the remote nodes for the time being and just run 1.5/1.6/1.7 on the Ubuntu node? Is that a valid work around? I would prefer to do that in the short term because getting this all this working doesn't seem like it's going to happen very quickly. Kristian 2012/11/26 Jason van Zyl ja...@tesla.io: Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A party which is not afraid of letting culture, business, and welfare go to ruin completely can be omnipotent for a while. -- Jakob Burckhardt
Re: Maven Core to Git
On 2012-11-26 15:59, Jason van Zyl wrote: On Nov 26, 2012, at 12:47 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: The problem is the asf builds running too often, and sometimes far too often and hence spamming the mailing lists substantially. Some of it has been solved, but there are still a few issues remaniing: As far as I can see there are two aspects of the issue: 1. A configuration error (or any kind of git corruption/inconsistency) on any of the git nodes would lead to the jenkins build running non-stop, generating mass amounts of spam on the notifications mailing lists. This issue has been solved as far as I can see; both by fixing the git configuration issue and by patching jenkins to fix the issue. This was https://issues.jenkins-ci.org/browse/JENKINS-15803. 2. There is a second issue where any kind of intermittent disconnect between the main jenkins instance and its node will trigger a rebuild because the master does not distinguish between a node being blank/reconfigured and simply unavaliable at the moment. This is basically happening by workspaceOffline returning true in https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L1437 I've been meaning to get Olivier to add logging in this code to find out what is actually happening. It seems to me like every single glitch in the asf network is causing rebuilds. And it would appear to be a quite unstable network So in practical terms it all needs to be released, and the servers all updated? There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Can we just turn off the remote nodes for the time being and just run 1.5/1.6/1.7 on the Ubuntu node? Is that a valid work around? I would prefer to do that in the short term because getting this all this working doesn't seem like it's going to happen very quickly. Which ubuntu node would that be? There are 5 of them and I think they are all slaves. Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? Kristian 2012/11/26 Jason van Zyl ja...@tesla.io: Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A party which is not afraid of letting culture, business, and welfare go to ruin completely can be omnipotent for a while. -- Jakob Burckhardt -- Dennis Lundberg - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
On Nov 26, 2012, at 10:41 AM, Dennis Lundberg denn...@apache.org wrote: On 2012-11-26 15:59, Jason van Zyl wrote: On Nov 26, 2012, at 12:47 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: The problem is the asf builds running too often, and sometimes far too often and hence spamming the mailing lists substantially. Some of it has been solved, but there are still a few issues remaniing: As far as I can see there are two aspects of the issue: 1. A configuration error (or any kind of git corruption/inconsistency) on any of the git nodes would lead to the jenkins build running non-stop, generating mass amounts of spam on the notifications mailing lists. This issue has been solved as far as I can see; both by fixing the git configuration issue and by patching jenkins to fix the issue. This was https://issues.jenkins-ci.org/browse/JENKINS-15803. 2. There is a second issue where any kind of intermittent disconnect between the main jenkins instance and its node will trigger a rebuild because the master does not distinguish between a node being blank/reconfigured and simply unavaliable at the moment. This is basically happening by workspaceOffline returning true in https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L1437 I've been meaning to get Olivier to add logging in this code to find out what is actually happening. It seems to me like every single glitch in the asf network is causing rebuilds. And it would appear to be a quite unstable network So in practical terms it all needs to be released, and the servers all updated? There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Can we just turn off the remote nodes for the time being and just run 1.5/1.6/1.7 on the Ubuntu node? Is that a valid work around? I would prefer to do that in the short term because getting this all this working doesn't seem like it's going to happen very quickly. Which ubuntu node would that be? There are 5 of them and I think they are all slaves. Just run on one machine for the matrix of JDKs we care about if this prevents the build explosions. Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. We have already haven't we? Apart from that I've done hundreds of releases out of Git and it works fine. But for core I will take responsibility if the buck needs to stop somewhere. I just want to get it moved over and work around any issues until the problems are solved. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? The release plugin works fine with Git. Kristian 2012/11/26 Jason van Zyl ja...@tesla.io: Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A party which is not afraid of letting culture, business, and welfare go to ruin completely can be omnipotent for a while. -- Jakob
Re: Maven Core to Git
On 2012-11-26 19:49, Jason van Zyl wrote: On Nov 26, 2012, at 10:41 AM, Dennis Lundberg denn...@apache.org wrote: On 2012-11-26 15:59, Jason van Zyl wrote: On Nov 26, 2012, at 12:47 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: The problem is the asf builds running too often, and sometimes far too often and hence spamming the mailing lists substantially. Some of it has been solved, but there are still a few issues remaniing: As far as I can see there are two aspects of the issue: 1. A configuration error (or any kind of git corruption/inconsistency) on any of the git nodes would lead to the jenkins build running non-stop, generating mass amounts of spam on the notifications mailing lists. This issue has been solved as far as I can see; both by fixing the git configuration issue and by patching jenkins to fix the issue. This was https://issues.jenkins-ci.org/browse/JENKINS-15803. 2. There is a second issue where any kind of intermittent disconnect between the main jenkins instance and its node will trigger a rebuild because the master does not distinguish between a node being blank/reconfigured and simply unavaliable at the moment. This is basically happening by workspaceOffline returning true in https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L1437 I've been meaning to get Olivier to add logging in this code to find out what is actually happening. It seems to me like every single glitch in the asf network is causing rebuilds. And it would appear to be a quite unstable network So in practical terms it all needs to be released, and the servers all updated? There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Can we just turn off the remote nodes for the time being and just run 1.5/1.6/1.7 on the Ubuntu node? Is that a valid work around? I would prefer to do that in the short term because getting this all this working doesn't seem like it's going to happen very quickly. Which ubuntu node would that be? There are 5 of them and I think they are all slaves. Just run on one machine for the matrix of JDKs we care about if this prevents the build explosions. Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. We have already haven't we? Apart from that I've done hundreds of releases out of Git and it works fine. But for core I will take responsibility if the buck needs to stop somewhere. I just want to get it moved over and work around any issues until the problems are solved. Like I said, I don't know the current status on this. It might very well be that we are there now. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? The release plugin works fine with Git. I'm sure it does. But for git beginners like myself who has not done a single release using git, we need a well documented release process for it. Kristian 2012/11/26 Jason van Zyl ja...@tesla.io: Kristian/Olivier, What exactly is the issue with switching over the core to Git? I only know vaguely what the reasoning is because I happened to wander into IRC one day. I also see the Jenkins issue[1] referred to in the Infra issue[2] about the conversion but it's not clear what's happening there or if it will be fixed anytime soon. What exactly is the problem? And why doesn't this behaviour exhibit itself in some of the other Git repos we have being built under Jenkins? I see tons of other projects using Git and Jenkins so can we workaround anything specific we're doing? [1]: https://issues.jenkins-ci.org/browse/JENKINS-15803 [2]: https://issues.apache.org/jira/browse/INFRA-5390 Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - You are never dedicated to something you have complete confidence in. No one is fanatically shouting that the sun is going to rise tomorrow. They know it is going to rise tomorrow. When people are fanatically dedicated to political or religious faiths or any other kind of dogmas or goals, it's always because these dogmas or goals are in doubt. -- Robert Pirzig, Zen and the Art of Motorcycle Maintenance - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder
Re: Maven Core to Git
There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode (https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
Kristian, can you ping me on irc tomorrow and maybe I can put some time into sorting the issues out (from the Jenkins side) On Monday, 26 November 2012, Kristian Rosenvold wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode ( https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354 ) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.org javascript:;
Re: Maven Core to Git
On the jenkins question, Why don't all the other Apache projects for which I'm on the dev list suffer from this? On the RM question, I don't think it's worth waiting. The core is the thing we release least often. We've run a release or two on the components we've already migrated, and whomever runs the first core release from git can be sure to add notes to the doc. On Mon, Nov 26, 2012 at 2:44 PM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: Kristian, can you ping me on irc tomorrow and maybe I can put some time into sorting the issues out (from the Jenkins side) On Monday, 26 November 2012, Kristian Rosenvold wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode ( https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354 ) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.org javascript:; - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
Agree. Shit happens, we'll deal with it. For now let's just lock it down to one machine. We can also just setup individual builds on the separate machines. No big deal. jvz On 2012-11-26, at 11:18 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode (https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
On 2012-11-26 20:18, Kristian Rosenvold wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode (https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Thanks Kristian, that's all I wanted to hear. Please go ahead and convert. Just remember to document any differences between svn and git along the way. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Dennis Lundberg - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Maven Core to Git
I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. If you're going to convert them, can you please keep them as individual repositories for now and I'd like to work with you through some use cases because I ran into some problems but you may have ways to work around them. I would prefer to merge them together later then assume it will work and have to undo it. On Nov 26, 2012, at 1:41 PM, Dennis Lundberg denn...@apache.org wrote: On 2012-11-26 20:18, Kristian Rosenvold wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode (https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Thanks Kristian, that's all I wanted to hear. Please go ahead and convert. Just remember to document any differences between svn and git along the way. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Dennis Lundberg - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - The modern conservative is engaged in one of man's oldest exercises in moral philosophy; that is, the search for a superior moral justification for selfishness. -- John Kenneth Galbraith
Re: Maven Core to Git
On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 If you're going to convert them, can you please keep them as individual repositories for now and I'd like to work with you through some use cases because I ran into some problems but you may have ways to work around them. I would prefer to merge them together later then assume it will work and have to undo it. On Nov 26, 2012, at 1:41 PM, Dennis Lundberg denn...@apache.org wrote: On 2012-11-26 20:18, Kristian Rosenvold wrote: There is also https://issues.jenkins-ci.org/browse/JENKINS-15367 which went into Jenkins 1.492 that was released yesterday, that may or may not be a factor in this depending who you talk to. Additionally, there is https://issues.jenkins-ci.org/browse/JENKINS-6604 which was fixed in remoting 4 days ago. According to the issue text it renders the slaves unusable. To which extent that makes the remote poller crash out is also unknown. But it's known that we have a *lot* of 6604 on the nodes, especially right after a restart. I am unsure what effect a single bad apple among the remote nodes has and to what extent we would detect it. Reading the jenkins code makes me quite sure that any build that is locked to 1 specific node will not encounter any of the network-down issues. I don't really understand how to get assignedNode ( https://github.com/jenkinsci/jenkins/blob/master/core/src/main/java/hudson/model/AbstractProject.java#L354 ) to be non-null but that should stop the random reallocations. I assume we could lock a few of the jobs down to given nodes and keep some open while we continue to track the problem ? Apart from these issues I proposed that we release a couple of our own products using git, before we move the core over to git. Just so that we have a good grasp of how a Maven release using git is done and get it properly documented. I'm not up to date on the progress here though. Would those of you that have done such releases please let us know? I believe all wagon, surefire and scm have all been released from git, so I think we're in the clear on /that/ particular aspect. Personally I think we should go ahead and convert core too. Thanks Kristian, that's all I wanted to hear. Please go ahead and convert. Just remember to document any differences between svn and git along the way. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- Dennis Lundberg - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - The modern conservative is engaged in one of man's oldest exercises in moral philosophy; that is, the search for a superior moral justification for selfishness. -- John Kenneth Galbraith -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier
Re: Maven Core to Git
On 27/11/2012, at 10:34 AM, Arnaud Héritier aherit...@gmail.com wrote: On Mon, Nov 26, 2012 at 11:20 PM, Jason van Zyl ja...@tesla.io wrote: I'm going to be working on the core for a few weeks. I am not convinced putting the ITs with the core is workable. I've tried it with a few scenarios and it's super confusing to me at least. +1 Agree - makes sense to keep them separate as they are often run against different versions of Maven. - Brett -- Brett Porter br...@apache.org http://brettporter.wordpress.com/ http://au.linkedin.com/in/brettporter http://twitter.com/brettporter - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Summary: Flipping Maven Core to Git
So we're almost at 4 weeks now. Does anyone know internally what's happening with: https://issues.apache.org/jira/browse/INFRA-5390 On Oct 13, 2012, at 7:14 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: After a very interesting discussion (which basically proves that community rocks IMO ;) I have created the following JIRA: https://issues.apache.org/jira/browse/INFRA-5390 Once this has gone r/w, we will merge in the existing M2 and IT repo. If we at a later stage decide that merging the IT's was undesirable, we can filter them back out into a separate repository. Kristian Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - believe nothing, no matter where you read it, or who has said it, not even if i have said it, unless it agrees with your own reason and your own common sense. -- Buddha
Re: Flipping Maven Core to Git
Given the general amount of attention that the community gives core changes, I think there is little reason to fear uncontrolled IT changes. Simply put; my fear of IT's *not* getting updated/maintained far exceeds my fear of an erronous update creeping in (which is why in surefire the IT's run by default with no option to turn them off other than -DskipTests ;) It seems to me we've covered all the important bases on this one; I'll summarize breiefly in a separate mail. Kristian 2012/10/12 Igor Fedorenko i...@ifedorenko.com Although I understand we can make the merged repository work, at least with some discipline from developers, I think merged source repository makes changes to the integration tests perceptually okay, which imho is rarely the case. I actually maintained fork of maven ITs for awhile, and personally I did not find it too much trouble. At least as long as I was only adding new ITs, not changing existing ones. Just my 2c. -- Regards, Igor On 12-10-12 3:47 AM, Kristian Rosenvold wrote: My mails came out sort-of in the wrong order, and I'm not sure if I agree with myself on splitting it changes - core changes. There is no technical mandate for this, not even a practical one since establishing mergeability from 2.2.1 - 3.X turned out to be simple. But as a personal work habit it's smart, because it allows you to change your mind easily. Let's say I want to fix MNG- on 3.X. I make an IT from a submitted sample project and I commit that on the 3.X branch. Now someone really wants that fixed back on 2.2.1, and it turns out the backport is real simple. All they need to do is cherry pick the IT and the core change back to 2.2.1. The IT always cherry-picks cleanly (more or less by definition due to the way we make IT's), but the core change does not, so the fix had to be re-applied on 2.2.1. When we merge 2.2.1 back to 3.X, git will automatically understand that the cherry-picked IT exists in the future and it won't trouble me with it. If the core change had been part of the IT commit I would lose this benefit, since the diffs become different (lol). (A diff that already exists in the history won't be reapplied, which is why it's so good to move *complete* commits backwards in time). Kristian 2012/10/12 Arnaud Héritier aherit...@gmail.com It may still be a good idea to do IT and core-change as two different commits If we merge core and ITs I think it will be important to do it (and thus to ask it to contributors - or we'll need to do it on our side). I always see few reasons to merge them and find the process a little bit complex (especially if the committer isn't a Git Guru) but I admit that I won't bit impacted by this choice as my core contributions were (and will probably continue to be) near to nothing. Even if I really support the Git migration I don't want that we decrease the ability to contribute (it is already a big issue) Arnaud On Fri, Oct 12, 2012 at 8:34 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Btw; this simplifies the overall workflow to: Check out the oldest version you want to support this feature and add your IT + change there. It may still be a good idea to do IT and core-change as two different commits Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/**maven-rc.githttps://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately
Summary: Flipping Maven Core to Git
After a very interesting discussion (which basically proves that community rocks IMO ;) I have created the following JIRA: https://issues.apache.org/jira/browse/INFRA-5390 Once this has gone r/w, we will merge in the existing M2 and IT repo. If we at a later stage decide that merging the IT's was undesirable, we can filter them back out into a separate repository. Kristian
Re: Flipping Maven Core to Git
I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of how day-to-day life of a branched workflow functions. When working in older history, it's usually wise to partition work into well-defined commits, because sometimes you might end up cherry-picking only parts of the work forward. When you're on trunk you usually don't need to worry about this stuff. {begin Theoretical Technical digression for the git-savy} Theoretically we should be able to make a single phoney merge commit that re-establishes the mergeability between the 2.2.1 and 3.0.X branch so that any subsequent changes we make to 2.2.1 can merge directly downstream. I have not done this in practice but I'm sure some others here have. I'm thinking something like this: git checkout maven3.0.X git merge -s ours origin/maven-2.2.1 Now of course I want to test this out before we proceed {end Theoretical Technical digression for the git-savy} Kristian 2012/10/11 Jason van Zyl ja...@tesla.io Say I have the following scenario: I'm working 3.1-S and creating some new ITs, and then I work on 2.2.x and create some ITs. What's the plan with making sure that we have one cohesive set of ITs that we can use across all versions? Just merge additions back into every branch? Because we would need to checkout two copies in order to be at one branch for the version of Maven you're working on and the branch where the ITs are. I understand wanting a rationalized fork and with unit tests I understand, but did you forks of large OSS projects include ITs like this? If that workflow is sorted out it seems fine. I just don't want it to be onerous to achieve the discipline of having on coherent set of tests that work across all versions of Maven. It's pretty easy to make that consistent right now. There's not a huge number of branches so merging additions back to each branch is pretty easy. Is that what you had in mind? On Oct 11, 2012, at 4:06 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: 2012/10/11 Arnaud Héritier aherit...@gmail.com Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? If you make any kind of change on a branch (bugfix/feature/whatever) I would expect to have test updates being done on that branch too, so the branch represents a complete diff of the change; both feature and tests. When the fix/feature is merged back into master the tests come along, if the feature is never merged to master, the tests stay in their branch - just like they should.
Re: Flipping Maven Core to Git
Btw; this simplifies the overall workflow to: Check out the oldest version you want to support this feature and add your IT + change there. It may still be a good idea to do IT and core-change as two different commits Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of how day-to-day life of a branched workflow functions. When working in older history, it's usually wise to partition work into well-defined commits, because sometimes you might end up cherry-picking only parts of the work forward. When you're on trunk you usually don't need to worry about this stuff. {begin Theoretical Technical digression for the git-savy} Theoretically we should be able to make a single phoney merge commit that re-establishes the mergeability between the 2.2.1 and 3.0.X branch so that any subsequent changes we make to 2.2.1 can merge directly downstream. I have not done this in practice but I'm sure some others here have. I'm thinking something like this: git checkout maven3.0.X git merge -s ours origin/maven-2.2.1 Now of course I want to test this out before we proceed {end Theoretical Technical digression for the git-savy} Kristian 2012/10/11 Jason van Zyl ja...@tesla.io Say I have the following scenario: I'm working 3.1-S and creating some new ITs, and then I work on 2.2.x and create some ITs. What's the plan with making sure that we have one cohesive set of ITs that we can use across all versions? Just merge additions back into every branch? Because we would need to checkout two copies in order to be at one branch for the version of Maven you're working on and the branch where the ITs are. I understand wanting a rationalized fork and with unit tests I understand, but did you forks of large OSS projects include ITs like this? If that workflow is sorted out it seems fine. I just don't want it to be onerous to achieve the discipline of having on coherent set of tests that work across all versions of Maven. It's pretty easy to make that consistent right now. There's not a huge number of branches so merging additions back to each branch is pretty easy. Is that what you had in mind? On Oct 11, 2012, at 4:06 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: 2012/10/11 Arnaud Héritier aherit...@gmail.com Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? If you make any kind of change on a branch (bugfix/feature/whatever) I would expect to have test
Re: Flipping Maven Core to Git
It may still be a good idea to do IT and core-change as two different commits If we merge core and ITs I think it will be important to do it (and thus to ask it to contributors - or we'll need to do it on our side). I always see few reasons to merge them and find the process a little bit complex (especially if the committer isn't a Git Guru) but I admit that I won't bit impacted by this choice as my core contributions were (and will probably continue to be) near to nothing. Even if I really support the Git migration I don't want that we decrease the ability to contribute (it is already a big issue) Arnaud On Fri, Oct 12, 2012 at 8:34 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Btw; this simplifies the overall workflow to: Check out the oldest version you want to support this feature and add your IT + change there. It may still be a good idea to do IT and core-change as two different commits Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of how day-to-day life of a branched workflow functions. When working in older history, it's usually wise to partition work into well-defined commits, because sometimes you might end up cherry-picking only parts of the work forward. When you're on trunk you usually don't need to worry about this stuff. {begin Theoretical Technical digression for the git-savy} Theoretically we should be able to make a single phoney merge commit that re-establishes the mergeability between the 2.2.1 and 3.0.X branch so that any subsequent changes we make to 2.2.1 can merge directly downstream. I have not done this in practice but I'm sure some others here have. I'm thinking something like this: git checkout maven3.0.X git merge -s ours origin/maven-2.2.1 Now of course I want to test this out before we proceed {end Theoretical Technical digression for the git-savy} Kristian 2012/10/11 Jason van Zyl ja...@tesla.io Say I have the following scenario: I'm working 3.1-S and creating some new ITs, and then I work on 2.2.x and create some ITs. What's the plan with making sure that we have one cohesive set of ITs that we can use across all versions? Just merge additions back into every branch? Because we would need to checkout two copies in order to be at one branch for the version of Maven you're working on and the branch where the ITs are. I understand wanting a rationalized fork and with unit tests I understand, but did you forks of large OSS projects include ITs like this? If that workflow is sorted out it seems fine. I just don't want it to be onerous to achieve the discipline of having on coherent set of tests that work across all
Re: Flipping Maven Core to Git
My mails came out sort-of in the wrong order, and I'm not sure if I agree with myself on splitting it changes - core changes. There is no technical mandate for this, not even a practical one since establishing mergeability from 2.2.1 - 3.X turned out to be simple. But as a personal work habit it's smart, because it allows you to change your mind easily. Let's say I want to fix MNG- on 3.X. I make an IT from a submitted sample project and I commit that on the 3.X branch. Now someone really wants that fixed back on 2.2.1, and it turns out the backport is real simple. All they need to do is cherry pick the IT and the core change back to 2.2.1. The IT always cherry-picks cleanly (more or less by definition due to the way we make IT's), but the core change does not, so the fix had to be re-applied on 2.2.1. When we merge 2.2.1 back to 3.X, git will automatically understand that the cherry-picked IT exists in the future and it won't trouble me with it. If the core change had been part of the IT commit I would lose this benefit, since the diffs become different (lol). (A diff that already exists in the history won't be reapplied, which is why it's so good to move *complete* commits backwards in time). Kristian 2012/10/12 Arnaud Héritier aherit...@gmail.com It may still be a good idea to do IT and core-change as two different commits If we merge core and ITs I think it will be important to do it (and thus to ask it to contributors - or we'll need to do it on our side). I always see few reasons to merge them and find the process a little bit complex (especially if the committer isn't a Git Guru) but I admit that I won't bit impacted by this choice as my core contributions were (and will probably continue to be) near to nothing. Even if I really support the Git migration I don't want that we decrease the ability to contribute (it is already a big issue) Arnaud On Fri, Oct 12, 2012 at 8:34 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Btw; this simplifies the overall workflow to: Check out the oldest version you want to support this feature and add your IT + change there. It may still be a good idea to do IT and core-change as two different commits Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of how day-to-day life of a branched workflow functions. When working in older history, it's usually wise to partition work into well-defined commits, because sometimes you might end up cherry-picking only parts of the work forward. When you're on trunk you usually don't need to worry about this stuff. {begin Theoretical
Re: Flipping Maven Core to Git
Although I understand we can make the merged repository work, at least with some discipline from developers, I think merged source repository makes changes to the integration tests perceptually okay, which imho is rarely the case. I actually maintained fork of maven ITs for awhile, and personally I did not find it too much trouble. At least as long as I was only adding new ITs, not changing existing ones. Just my 2c. -- Regards, Igor On 12-10-12 3:47 AM, Kristian Rosenvold wrote: My mails came out sort-of in the wrong order, and I'm not sure if I agree with myself on splitting it changes - core changes. There is no technical mandate for this, not even a practical one since establishing mergeability from 2.2.1 - 3.X turned out to be simple. But as a personal work habit it's smart, because it allows you to change your mind easily. Let's say I want to fix MNG- on 3.X. I make an IT from a submitted sample project and I commit that on the 3.X branch. Now someone really wants that fixed back on 2.2.1, and it turns out the backport is real simple. All they need to do is cherry pick the IT and the core change back to 2.2.1. The IT always cherry-picks cleanly (more or less by definition due to the way we make IT's), but the core change does not, so the fix had to be re-applied on 2.2.1. When we merge 2.2.1 back to 3.X, git will automatically understand that the cherry-picked IT exists in the future and it won't trouble me with it. If the core change had been part of the IT commit I would lose this benefit, since the diffs become different (lol). (A diff that already exists in the history won't be reapplied, which is why it's so good to move *complete* commits backwards in time). Kristian 2012/10/12 Arnaud Héritier aherit...@gmail.com It may still be a good idea to do IT and core-change as two different commits If we merge core and ITs I think it will be important to do it (and thus to ask it to contributors - or we'll need to do it on our side). I always see few reasons to merge them and find the process a little bit complex (especially if the committer isn't a Git Guru) but I admit that I won't bit impacted by this choice as my core contributions were (and will probably continue to be) near to nothing. Even if I really support the Git migration I don't want that we decrease the ability to contribute (it is already a big issue) Arnaud On Fri, Oct 12, 2012 at 8:34 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Btw; this simplifies the overall workflow to: Check out the oldest version you want to support this feature and add your IT + change there. It may still be a good idea to do IT and core-change as two different commits Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com I could not resist, I made this repo and it works well: git clone https://github.com/krosenvold/maven-rc.git It has reconstructed mergeability between 2.2.1 and 3.X, and the IT's were added at 2.2.1 and merged up to 3.X. Which means we have a straight merge based workflow from 2.2.1 and on. Pre 2.2.1 versions are not supported for this mode of operation. Note: Older tags (pre. 2.2.0) will have to be reconstructed by hand once we flip. Also; the pom.xml and the site in this merged repo contain only the core pom (changes from core-its pom.xml and site will have to be merged by hand) Btw: In this repo, you can checkout trunk (3.X and do git diff ...2.2.X and see what changes have been added to 2.2.X. Note three ...'s) Kristian 2012/10/12 Kristian Rosenvold kristian.rosenv...@gmail.com NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of
Re: Flipping Maven Core to Git
So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
Kristian, wait a bit. What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! LieGrue, strub - Original Message - From: Kristian Rosenvold kristian.rosenv...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Sent: Thursday, October 11, 2012 8:35 AM Subject: Re: Flipping Maven Core to Git Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
2012/10/11 Mark Struberg strub...@yahoo.de: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.org javascript:;
Re: Flipping Maven Core to Git
$ ls -1b maven-3/ README.bootstrap.txt README.md README.txt apache-maven build.xml doap_Maven.rdf maven-aether-provider maven-ant-tasks-2.1.1.jar maven-artifact maven-compat maven-core maven-embedder maven-model maven-model-builder maven-plugin-api maven-repository-metadata maven-settings maven-settings-builder maven.iml pom.xml src target $ ls -1b core-its/ README core-it-suite core-it-support pom.xml src I really think they could just end up being side by side (without root folders for each) and merge the root pom by hand, adding the two core-it-* modules and setting up a run-its profile like the rest of our projects. I think the argument for doing this is that it increases the focus on the existence of the core it suite. I really see no significant downsides ? Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.orgjavascript:; -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier
Re: Flipping Maven Core to Git
As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.orgjavascript:; -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
Initially I thought Arnaud's argument for being able to run against a different maven version was a killer; we needed to keep them separate. After giving it some more thought I'm not sure it's a good argument, we don't lose the ability to do so if we merge the trees (it's just one of many possible flags to use when running the its, all available in the docs) So I'm still marginally in favour of merging the trees, if only to increas focus on the IT's. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
As far as it is always possible (and easy) to do (and to setup on CI side as it will be the first one to run them) I'll follow you. Arnaud On Thu, Oct 11, 2012 at 9:56 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Initially I thought Arnaud's argument for being able to run against a different maven version was a killer; we needed to keep them separate. After giving it some more thought I'm not sure it's a good argument, we don't lose the ability to do so if we merge the trees (it's just one of many possible flags to use when running the its, all available in the docs) So I'm still marginally in favour of merging the trees, if only to increas focus on the IT's. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier
Re: Flipping Maven Core to Git
Are we then also saying that we merge the maven-2 source into this git repo as well (if we convert that to git)? /Anders On Thu, Oct 11, 2012 at 10:01 AM, Arnaud Héritier aherit...@gmail.com wrote: As far as it is always possible (and easy) to do (and to setup on CI side as it will be the first one to run them) I'll follow you. Arnaud On Thu, Oct 11, 2012 at 9:56 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Initially I thought Arnaud's argument for being able to run against a different maven version was a killer; we needed to keep them separate. After giving it some more thought I'm not sure it's a good argument, we don't lose the ability to do so if we merge the trees (it's just one of many possible flags to use when running the its, all available in the docs) So I'm still marginally in favour of merging the trees, if only to increas focus on the IT's. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
Why too much rush ? Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition
Re: Flipping Maven Core to Git
we need to revert/fix the log commit anyway for now as we don't yet have the classloader isolation code. LieGrue, strub - Original Message - From: Olivier Lamy ol...@apache.org To: Maven Developers List dev@maven.apache.org Cc: Sent: Thursday, October 11, 2012 10:19 AM Subject: Re: Flipping Maven Core to Git Why too much rush ? Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache
Re: Flipping Maven Core to Git
On Oct 11, 2012, at 12:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? That's not what I said. It is possible the run the ITs while in subversion while the core is in Git. Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - We know what we are, but know not what we may be. -- Shakespeare
Re: Flipping Maven Core to Git
On Oct 11, 2012, at 2:41 AM, Mark Struberg strub...@yahoo.de wrote: Kristian, wait a bit. What if we first merge the 2 repos into 1 git repo? -1 They are separate for the specific reason that the body of ITs be independent from the core and easily run against multiple versions of Maven. Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. No one has ever forgotten about them since they were first written. Why is that a concern? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! LieGrue, strub - Original Message - From: Kristian Rosenvold kristian.rosenv...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Sent: Thursday, October 11, 2012 8:35 AM Subject: Re: Flipping Maven Core to Git Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl
Re: Flipping Maven Core to Git
I just merged the maven-2 and the maven-3 git repo back together using something like this: git clone http://git.apache.org/maven-3.git cd maven-3 git remote add m2 http://git.apache.org/maven-2.git git fetch m2 And it works perfectly; you can even run git merge-base maven-2.2.x trunk to find the branch point ;) They share common history in the existing git clones. If we move the maven-3 repo I think we might call it maven and just re-add the maven-2 history. Kristian 2012/10/11 Anders Hammar and...@hammar.net: Are we then also saying that we merge the maven-2 source into this git repo as well (if we convert that to git)? /Anders On Thu, Oct 11, 2012 at 10:01 AM, Arnaud Héritier aherit...@gmail.com wrote: As far as it is always possible (and easy) to do (and to setup on CI side as it will be the first one to run them) I'll follow you. Arnaud On Thu, Oct 11, 2012 at 9:56 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Initially I thought Arnaud's argument for being able to run against a different maven version was a killer; we needed to keep them separate. After giving it some more thought I'm not sure it's a good argument, we don't lose the ability to do so if we merge the trees (it's just one of many possible flags to use when running the its, all available in the docs) So I'm still marginally in favour of merging the trees, if only to increas focus on the IT's. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.orgjavascript:; -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - believe nothing, no matter where you read it, or who has said it, not even if i have said it, unless it agrees with your own reason and your own common sense. -- Buddha
Re: Flipping Maven Core to Git
I would prefer to have one maven-core git repo with various branches as we don't have the SVN constraint that justified to have several entry points instead of branches. On Thu, Oct 11, 2012 at 10:58 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: I just merged the maven-2 and the maven-3 git repo back together using something like this: git clone http://git.apache.org/maven-3.git cd maven-3 git remote add m2 http://git.apache.org/maven-2.git git fetch m2 And it works perfectly; you can even run git merge-base maven-2.2.x trunk to find the branch point ;) They share common history in the existing git clones. If we move the maven-3 repo I think we might call it maven and just re-add the maven-2 history. Kristian 2012/10/11 Anders Hammar and...@hammar.net: Are we then also saying that we merge the maven-2 source into this git repo as well (if we convert that to git)? /Anders On Thu, Oct 11, 2012 at 10:01 AM, Arnaud Héritier aherit...@gmail.com wrote: As far as it is always possible (and easy) to do (and to setup on CI side as it will be the first one to run them) I'll follow you. Arnaud On Thu, Oct 11, 2012 at 9:56 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Initially I thought Arnaud's argument for being able to run against a different maven version was a killer; we needed to keep them separate. After giving it some more thought I'm not sure it's a good argument, we don't lose the ability to do so if we merge the trees (it's just one of many possible flags to use when running the its, all available in the docs) So I'm still marginally in favour of merging the trees, if only to increas focus on the IT's. Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier
Re: Flipping Maven Core to Git
On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van
Re: Flipping Maven Core to Git
2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language
Re: Flipping Maven Core to Git
I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core
Re: Flipping Maven Core to Git
I think the key argument for adding the IT's to the maven core repo is that it makes it so much easier to maintain high quality forks, a scenario I think we should fully embrace since it's also a gateway point for high-quality contributions. Having maintained quite extensive forks of several *large* OSS frameworks for some years in git, I think the forker has a goal of knowing exactly what the difference is between her fork and the mainline. As the mainline evolves, I think it's crucial for the forker to be able to evolve. Keeping the IT's together with the code makes it a lot easier to keep track of which changes a fork actually contain, since the branch effectively contains not only core changes but also IT changes. The IT's determine how maven-like your fork is. The forker can maintain a clear notion of this by changing his forked IT's too. You have this great tool for asserting the quality of the work/fork and you decide to make it twice as hard to use by keeping them in separate repos. Most of the stuff you refer to is simply the way things have been done up to now. The coupling concern is the only one that I feel has some technical backing. A different technical argument is that if we add the IT's to the core repo, we could choose to run some of them as part of a regular clean install build. We have done this with great success on some in-house projects; tag the IT's with JUnit @Category minimal test suite and have that run with the main build. So I say m2+m3+it's in one repo is a neat solution. The ability to test other mavens will be there no matter what ;) Kristian 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.orgjavascript:; -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO
Re: Flipping Maven Core to Git
OK, you have won me over. Lets keep them in one repo. We can do some work down the line to make the ITs a separately releasable bundle and that way make it easy to run the latest ITs against older versions and vice versa better to get the git repo right from the start rather than rewrite history to get a proper history. On 11 October 2012 10:40, Kristian Rosenvold kristian.rosenv...@gmail.comwrote: I think the key argument for adding the IT's to the maven core repo is that it makes it so much easier to maintain high quality forks, a scenario I think we should fully embrace since it's also a gateway point for high-quality contributions. Having maintained quite extensive forks of several *large* OSS frameworks for some years in git, I think the forker has a goal of knowing exactly what the difference is between her fork and the mainline. As the mainline evolves, I think it's crucial for the forker to be able to evolve. Keeping the IT's together with the code makes it a lot easier to keep track of which changes a fork actually contain, since the branch effectively contains not only core changes but also IT changes. The IT's determine how maven-like your fork is. The forker can maintain a clear notion of this by changing his forked IT's too. You have this great tool for asserting the quality of the work/fork and you decide to make it twice as hard to use by keeping them in separate repos. Most of the stuff you refer to is simply the way things have been done up to now. The coupling concern is the only one that I feel has some technical backing. A different technical argument is that if we add the IT's to the core repo, we could choose to run some of them as part of a regular clean install build. We have done this with great success on some in-house projects; tag the IT's with JUnit @Category minimal test suite and have that run with the main build. So I say m2+m3+it's in one repo is a neat solution. The ability to test other mavens will be there no matter what ;) Kristian 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr
Re: Flipping Maven Core to Git
Fair enough. I can't think of a reason not to try it. In Git it's easy to change anything if it's not working out in practice. On Oct 11, 2012, at 5:40 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: I think the key argument for adding the IT's to the maven core repo is that it makes it so much easier to maintain high quality forks, a scenario I think we should fully embrace since it's also a gateway point for high-quality contributions. Having maintained quite extensive forks of several *large* OSS frameworks for some years in git, I think the forker has a goal of knowing exactly what the difference is between her fork and the mainline. As the mainline evolves, I think it's crucial for the forker to be able to evolve. Keeping the IT's together with the code makes it a lot easier to keep track of which changes a fork actually contain, since the branch effectively contains not only core changes but also IT changes. The IT's determine how maven-like your fork is. The forker can maintain a clear notion of this by changing his forked IT's too. You have this great tool for asserting the quality of the work/fork and you decide to make it twice as hard to use by keeping them in separate repos. Most of the stuff you refer to is simply the way things have been done up to now. The coupling concern is the only one that I feel has some technical backing. A different technical argument is that if we add the IT's to the core repo, we could choose to run some of them as part of a regular clean install build. We have done this with great success on some in-house projects; tag the IT's with JUnit @Category minimal test suite and have that run with the main build. So I say m2+m3+it's in one repo is a neat solution. The ability to test other mavens will be there no matter what ;) Kristian 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org javascript:; For additional commands, e-mail: dev-h...@maven.apache.orgjavascript:; -- - Arnaud Héritier 06-89-76-64-24 http://aheritier.net Mail/GTalk: aherit...@gmail.com Twitter/Skype : aheritier
RE: Flipping Maven Core to Git
Hi Oliverwhat are the advantages of moving from tried and true (works everytime) workhorse of svn to git version control ? thanks, From: ol...@apache.org Date: Thu, 11 Oct 2012 11:31:04 +0200 Subject: Re: Flipping Maven Core to Git To: dev@maven.apache. 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian
Re: Flipping Maven Core to Git
On Thu, Oct 11, 2012 at 7:43 AM, Martin Gainty mgai...@hotmail.com wrote: Hi Oliverwhat are the advantages of moving from tried and true (works everytime) workhorse of svn to git version control ? We already discussed and voted this question. let's not do it again. thanks, From: ol...@apache.org Date: Thu, 11 Oct 2012 11:31:04 +0200 Subject: Re: Flipping Maven Core to Git To: dev@maven.apache. 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need
Re: Flipping Maven Core to Git
2012/10/11 Jason van Zyl ja...@tesla.io: I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. nice. BTW some its are failing too: https://builds.apache.org/view/M-R/view/Maven/job/core-integration-testing-maven-3/lastCompletedBuild/testReport/ On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in
Re: Flipping Maven Core to Git
Yup, the ITs don't seem to run when changes are made. Are they manually triggered? I waited last night for a while but nothing was happening. On Oct 11, 2012, at 8:30 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. nice. BTW some its are failing too: https://builds.apache.org/view/M-R/view/Maven/job/core-integration-testing-maven-3/lastCompletedBuild/testReport/ On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we
Re: Flipping Maven Core to Git
Is there actually a MAC in the test grid as I see it's grayed out and hasn't run in over a year? Any particular reason? On Oct 11, 2012, at 8:30 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. nice. BTW some its are failing too: https://builds.apache.org/view/M-R/view/Maven/job/core-integration-testing-maven-3/lastCompletedBuild/testReport/ On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off
Re: Flipping Maven Core to Git
2012/10/11 Jason van Zyl ja...@tesla.io: Yup, the ITs don't seem to run when changes are made. Are they manually triggered? I waited last night for a while but nothing was happening. there are triggered by changes in http://svn.apache.org/repos/asf/maven/maven-3/trunk or in core its source tree. If I correctly read here https://builds.apache.org/view/M-R/view/Maven/ they runed around 13H ago (depending on nodes some are not used a lot by other asf projects) 13H ago is very similar to your last commit time On Oct 11, 2012, at 8:30 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. nice. BTW some its are failing too: https://builds.apache.org/view/M-R/view/Maven/job/core-integration-testing-maven-3/lastCompletedBuild/testReport/ On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty... /Anders On Thu, Oct 11, 2012 at 6:18 AM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be
Re: Flipping Maven Core to Git
I have always a doubt about ITs maintenance if they are in the same repo as the core itself For Its we are not really releasing them nor we managed their versions. we took the trunk (or the master nowadays in git) and we launched them on a given maven version Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? WDYT ? On Thu, Oct 11, 2012 at 11:51 AM, Jason van Zyl ja...@tesla.io wrote: Fair enough. I can't think of a reason not to try it. In Git it's easy to change anything if it's not working out in practice. On Oct 11, 2012, at 5:40 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: I think the key argument for adding the IT's to the maven core repo is that it makes it so much easier to maintain high quality forks, a scenario I think we should fully embrace since it's also a gateway point for high-quality contributions. Having maintained quite extensive forks of several *large* OSS frameworks for some years in git, I think the forker has a goal of knowing exactly what the difference is between her fork and the mainline. As the mainline evolves, I think it's crucial for the forker to be able to evolve. Keeping the IT's together with the code makes it a lot easier to keep track of which changes a fork actually contain, since the branch effectively contains not only core changes but also IT changes. The IT's determine how maven-like your fork is. The forker can maintain a clear notion of this by changing his forked IT's too. You have this great tool for asserting the quality of the work/fork and you decide to make it twice as hard to use by keeping them in separate repos. Most of the stuff you refer to is simply the way things have been done up to now. The coupling concern is the only one that I feel has some technical backing. A different technical argument is that if we add the IT's to the core repo, we could choose to run some of them as part of a regular clean install build. We have done this with great success on some in-house projects; tag the IT's with JUnit @Category minimal test suite and have that run with the main build. So I say m2+m3+it's in one repo is a neat solution. The ability to test other mavens will be there no matter what ;) Kristian 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact
Re: Flipping Maven Core to Git
On Oct 11, 2012, at 8:42 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: Yup, the ITs don't seem to run when changes are made. Are they manually triggered? I waited last night for a while but nothing was happening. there are triggered by changes in http://svn.apache.org/repos/asf/maven/maven-3/trunk or in core its source tree. If I correctly read here https://builds.apache.org/view/M-R/view/Maven/ they runed around 13H ago (depending on nodes some are not used a lot by other asf projects) 13H ago is very similar to your last commit time I waited a while, and if both systems are on UTC the difference between the checkin and the build is an hour. Checkin: Wed Oct 10 22:48:14 2012 UTC (14 hours, 8 minutes ago) http://svn.apache.org/viewvc?view=revisionrevision=1396843 Build: #203 (Oct 10, 2012 11:07:47 PM) (13 hours ago) Is it tied to LDAP as I tried to login to trigger the build manually but I couldn't log in. On Oct 11, 2012, at 8:30 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: I can look at both of those, maybe we can figure something out in the simple implementation to make it look better. Ant thing looks easy enough to fix. nice. BTW some its are failing too: https://builds.apache.org/view/M-R/view/Maven/job/core-integration-testing-maven-3/lastCompletedBuild/testReport/ On Oct 11, 2012, at 5:31 AM, Olivier Lamy ol...@apache.org wrote: 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 4:19 AM, Olivier Lamy ol...@apache.org wrote: Why too much rush ? I just find it more convenient to have everything in Git. Makes the testing cycle more pleasant and the git svn workflow isn't something I personally enjoy. I'm working on the core and it's nothing more than a preference. Perso, I will be happy to see similar rush on writing test/doc the stuff committed 1,5 month ago !.(despite a number of ping) AFAIK (maybe I miss something) using @Inject is not so trivial because we need to generate the sisu index. Which is very similar to what we do currently with plexus plugins which generate metadata. So what is the added value for users or plugin developers (except using the last à la mode @Inject). I will be happy to hear the added values (performance ? etc .. ?) I'll answer these in another post. With the documentation, that will be easier. regarding last commit regarding logs in trunk, I tested that and the artifact transfer progress(download/upload) is now (IMHO) very crappy !! I'm not sure technically what very crappy means. You mean the way it visually looks? Yup visible changes for users will be a tons of lines in the console. Maybe it's just matter of log layout to configure (As I didn't have a look I don't know) Something else when building core with ant bootstrap (sure not really important :-)): process-classes: [echo] Using plexus version 1.5.5 [java] SLF4J: Failed to load class org.slf4j.impl.StaticLoggerBinder. [java] SLF4J: Defaulting to no-operation (NOP) logger implementation [java] SLF4J: See http://www.slf4j.org/codes.html#StaticLoggerBinder for further details. [INFO] Building SCM Publish Maven Plugin 1.0-SNAPSHOT [INFO] [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom [INFO] 4/16 KB [INFO] 8/16 KB [INFO] 12/16 KB [INFO] 16/16 KB [INFO] 16/16 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugin-tools/maven-plugin-tools/3.1/maven-plugin-tools-3.1.pom (16 KB at 9.2 KB/sec) [INFO] Downloading: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar [INFO] 4/47 KB [INFO] 8/47 KB [INFO] 12/47 KB [INFO] 16/47 KB [INFO] 20/47 KB [INFO] 24/47 KB [INFO] 28/47 KB [INFO] 31/47 KB [INFO] 35/47 KB [INFO] 38/47 KB [INFO] 42/47 KB [INFO] 46/47 KB [INFO] 47/47 KB [INFO] [INFO] Downloaded: https://archiva-repository.apache.org/archiva/repository/public/org/apache/maven/plugins/maven-plugin-plugin/3.1/maven-plugin-plugin-3.1.jar (47 KB at 45.8 KB/sec) Sorry I didn't investigate more on the reason. 2012/10/11 Kristian Rosenvold kristian.rosenv...@gmail.com: Jason ran the tests based off the git repo and they were fine. There is no problem (I've had my fair share of troubles with the core it's so it's probably been a problem between chair and keyboard). Unless someone objects I'll file a INFRA jira to port m3-core + IT's sometime later today. Kristian 2012/10/11 Anders Hammar and...@hammar.net: So, what is the solution to fixing the ITs that depend on an empty folder? Put some dummy file there? But then the folder will not be empty...
Re: Flipping Maven Core to Git
To that end we should probably also try to collect any specific ITs for plugins that are currently in the IT suite and put them with their respective plugins. I think the idea of knowing what you're forking is wise. It would be really hard for someone to know we have tests for specific plugins in the ITs and that they should be run to validate a particular plugin. On Oct 11, 2012, at 5:40 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: I think the key argument for adding the IT's to the maven core repo is that it makes it so much easier to maintain high quality forks, a scenario I think we should fully embrace since it's also a gateway point for high-quality contributions. Having maintained quite extensive forks of several *large* OSS frameworks for some years in git, I think the forker has a goal of knowing exactly what the difference is between her fork and the mainline. As the mainline evolves, I think it's crucial for the forker to be able to evolve. Keeping the IT's together with the code makes it a lot easier to keep track of which changes a fork actually contain, since the branch effectively contains not only core changes but also IT changes. The IT's determine how maven-like your fork is. The forker can maintain a clear notion of this by changing his forked IT's too. You have this great tool for asserting the quality of the work/fork and you decide to make it twice as hard to use by keeping them in separate repos. Most of the stuff you refer to is simply the way things have been done up to now. The coupling concern is the only one that I feel has some technical backing. A different technical argument is that if we add the IT's to the core repo, we could choose to run some of them as part of a regular clean install build. We have done this with great success on some in-house projects; tag the IT's with JUnit @Category minimal test suite and have that run with the main build. So I say m2+m3+it's in one repo is a neat solution. The ability to test other mavens will be there no matter what ;) Kristian 2012/10/11 Jason van Zyl ja...@tesla.io: On Oct 11, 2012, at 3:44 AM, Mark Struberg strub...@yahoo.de wrote: As I see it after the feedback there are 2 main arguments: * maintaining the ITs during maven-core development. People currently tend to forget adding ITs for the stuff they change. This might be easier if the ITs are contained in the same repo. * being able to run the ITs onto old maven releases. That's for sure easier if we keep em separated. Any other arguments to add to the list? - they do not have the same lifecycle - the are designed to be separate - coupling will most certainly happen if people work on them simultaneously I'm perfectly fine either way. It's just that we can change this easily right now and it might be harder to do later. LieGrue, strub - Original Message - From: Arnaud Héritier aherit...@gmail.com To: Maven Developers List dev@maven.apache.org Cc: Mark Struberg strub...@yahoo.de Sent: Thursday, October 11, 2012 9:32 AM Subject: Re: Flipping Maven Core to Git Like Stephen I would prefer to keep them separated. They have a different lifecycle as we should be (we are ?) able to run ITs against various versions of Maven and we take care to have flags to enable/disable some tests. I see no advantage to merge them For me the need to reduce the number of repositories is like the need to reduce the number of Jira projects. It's only a sysadmin constraint and it is against the spirit of these tools. Arnaud On Thu, Oct 11, 2012 at 9:07 AM, Stephen Connolly stephen.alan.conno...@gmail.com wrote: On Thursday, 11 October 2012, Kristian Rosenvold wrote: 2012/10/11 Mark Struberg strub...@yahoo.de javascript:;: What if we first merge the 2 repos into 1 git repo? Imo maven-core and the ITs must fit together! Having the ITs in a separate repo will make people forget about them. None of them are big, we could easily merge them; conceptually they belong together. It would seem like they should be merged with the same source root. I assume the best way to do this would simply be to flip m3 to git, then just add the complete history of core-its and then just merge them on trunk... ? Also for a maven-core release it isn't a problem if the ITs get tagged. I actually would even really appreciate that fact! Each core iIT is tagged with the initial maven version which it is valid at (using a homebrew replacement of JUnit assumptions), which means the full IT suite is self-configuring wrt which maven version it is being run against. No need to tag it really. I would actually prefer that the it suite is separate as it is version independent. But I don't feel strongly either way Kristian - To unsubscribe, e-mail: dev-unsubscr
Re: Flipping Maven Core to Git
2012/10/11 Arnaud Héritier aherit...@gmail.com Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? If you make any kind of change on a branch (bugfix/feature/whatever) I would expect to have test updates being done on that branch too, so the branch represents a complete diff of the change; both feature and tests. When the fix/feature is merged back into master the tests come along, if the feature is never merged to master, the tests stay in their branch - just like they should. This is a healthy mode of operation; if you look at the branches in my surefire github repo (https://github.com/krosenvold/maven-surefire) I have like 15 different branches. Some of them are specific JIRAs and may contain testcases and partial fixes. Now I will start pushing these to the ASF git repo instead, since they represent work on an issue that is not fully fixed; maybe just a testcase or a test project. I think it's important we acknowledge the existence/intention of such branches on the corresponding JIRA when we make such feature branches in the official repo. If I just push it to my personal github repo it can be as messy as I prefer, but when I push it to ASF I'd prefer it to have reasonably well-defined responsibility (like test-case, suggested fix or similar) and some value to others. If it's just my personal doodles I'll keep them in my personal github. As for branches that never get merged back; well that's life and usually there's a story behind that and we tell it in jira. Kristian
Re: Flipping Maven Core to Git
Say I have the following scenario: I'm working 3.1-S and creating some new ITs, and then I work on 2.2.x and create some ITs. What's the plan with making sure that we have one cohesive set of ITs that we can use across all versions? Just merge additions back into every branch? Because we would need to checkout two copies in order to be at one branch for the version of Maven you're working on and the branch where the ITs are. I understand wanting a rationalized fork and with unit tests I understand, but did you forks of large OSS projects include ITs like this? If that workflow is sorted out it seems fine. I just don't want it to be onerous to achieve the discipline of having on coherent set of tests that work across all versions of Maven. It's pretty easy to make that consistent right now. There's not a huge number of branches so merging additions back to each branch is pretty easy. Is that what you had in mind? On Oct 11, 2012, at 4:06 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: 2012/10/11 Arnaud Héritier aherit...@gmail.com Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? If you make any kind of change on a branch (bugfix/feature/whatever) I would expect to have test updates being done on that branch too, so the branch represents a complete diff of the change; both feature and tests. When the fix/feature is merged back into master the tests come along, if the feature is never merged to master, the tests stay in their branch - just like they should. This is a healthy mode of operation; if you look at the branches in my surefire github repo (https://github.com/krosenvold/maven-surefire) I have like 15 different branches. Some of them are specific JIRAs and may contain testcases and partial fixes. Now I will start pushing these to the ASF git repo instead, since they represent work on an issue that is not fully fixed; maybe just a testcase or a test project. I think it's important we acknowledge the existence/intention of such branches on the corresponding JIRA when we make such feature branches in the official repo. If I just push it to my personal github repo it can be as messy as I prefer, but when I push it to ASF I'd prefer it to have reasonably well-defined responsibility (like test-case, suggested fix or similar) and some value to others. If it's just my personal doodles I'll keep them in my personal github. As for branches that never get merged back; well that's life and usually there's a story behind that and we tell it in jira. Kristian Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - In short, man creates for himself a new religion of a rational and technical order to justify his work and to be justified in it. -- Jacques Ellul, The Technological Society
Re: Flipping Maven Core to Git
NIce use case, this is where things get interesting ;) Initially I'll just add that this is one of those workflows that might be easier with a separate IT repo. This is how the flow would be in a merged single repo: This workflow has two distinct modes of operation: 1. You always check out the oldest appropriate branch to make the IT. So if the IT should apply to 2.2.1 you should check that out. (If we need to make IT's that go even further back I think it might be easiest to just keep the IT's in a separate repo since it gets really hard to actually *make* that repo.) 2 A) You are working in a properly branched structure that is mergeable ; i.e. what we might do between 3.0.X and 3.1 or maybe more appropriately between 3.X and 4.X. Make all the changes your heart desires, in any kind of commit sequence you prefer. The downstream merge to the subsequent version will fix it all. 2 B) You are working i an unmergable branch structure (2.2.1-3.0) Add the IT as a distinct commit on the 2.2.1 branch, make sure any bug-fix you do is done as a separate commit from the IT. When done, you will need to check out the 3.0.X branch and cherry pick the commit with the IT. From 3.0.X on it will merge properly. cherry-picking is a manual operation and if you forget it, the IT will be lost on 2.2.1 While this all may seem a bit complex, it's basically the nuts bolts of how day-to-day life of a branched workflow functions. When working in older history, it's usually wise to partition work into well-defined commits, because sometimes you might end up cherry-picking only parts of the work forward. When you're on trunk you usually don't need to worry about this stuff. {begin Theoretical Technical digression for the git-savy} Theoretically we should be able to make a single phoney merge commit that re-establishes the mergeability between the 2.2.1 and 3.0.X branch so that any subsequent changes we make to 2.2.1 can merge directly downstream. I have not done this in practice but I'm sure some others here have. I'm thinking something like this: git checkout maven3.0.X git merge -s ours origin/maven-2.2.1 Now of course I want to test this out before we proceed {end Theoretical Technical digression for the git-savy} Kristian 2012/10/11 Jason van Zyl ja...@tesla.io Say I have the following scenario: I'm working 3.1-S and creating some new ITs, and then I work on 2.2.x and create some ITs. What's the plan with making sure that we have one cohesive set of ITs that we can use across all versions? Just merge additions back into every branch? Because we would need to checkout two copies in order to be at one branch for the version of Maven you're working on and the branch where the ITs are. I understand wanting a rationalized fork and with unit tests I understand, but did you forks of large OSS projects include ITs like this? If that workflow is sorted out it seems fine. I just don't want it to be onerous to achieve the discipline of having on coherent set of tests that work across all versions of Maven. It's pretty easy to make that consistent right now. There's not a huge number of branches so merging additions back to each branch is pretty easy. Is that what you had in mind? On Oct 11, 2012, at 4:06 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: 2012/10/11 Arnaud Héritier aherit...@gmail.com Now let's say that someone wants to solve a problem on a maintenance branch (thus not master) With git it will checkout this branch to work on the fix, but if in // he wants to add an IT it will add it on this branch (not on master where we should have all ITs ?) How many chance that one day we forgot to merge back ITs from all branches in master ? If you make any kind of change on a branch (bugfix/feature/whatever) I would expect to have test updates being done on that branch too, so the branch represents a complete diff of the change; both feature and tests. When the fix/feature is merged back into master the tests come along, if the feature is never merged to master, the tests stay in their branch - just like they should. This is a healthy mode of operation; if you look at the branches in my surefire github repo (https://github.com/krosenvold/maven-surefire) I have like 15 different branches. Some of them are specific JIRAs and may contain testcases and partial fixes. Now I will start pushing these to the ASF git repo instead, since they represent work on an issue that is not fully fixed; maybe just a testcase or a test project. I think it's important we acknowledge the existence/intention of such branches on the corresponding JIRA when we make such feature branches in the official repo. If I just push it to my personal github repo it can be as messy as I prefer, but when I push it to ASF I'd prefer it to have reasonably well-defined responsibility (like test-case, suggested fix or similar) and some value to
Converting Maven Core to Git
Hi, Where are the scripts that were used to convert the other repos to Git? I'm going to spend tomorrow trying to convert Maven core over to Git as trying to finish the other work I'm trying to do and using Subversion is getting in my way so I'd just like to be done with moving the core over to Git. Pure selfish motivation. Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - What matters is not ideas, but the people who have them. Good people can fix bad ideas, but good ideas can't save bad people. -- Paul Graham
Re: Converting Maven Core to Git
The repo is already ported, it's just a matter having infra flip some switches. ISTR there are some issues that need to be fixed in the core it's when running from a git repo Those issues can be fixed now ;) Kristian Den 10. okt. 2012 kl. 18:28 skrev Jason van Zyl ja...@tesla.io: Hi, Where are the scripts that were used to convert the other repos to Git? I'm going to spend tomorrow trying to convert Maven core over to Git as trying to finish the other work I'm trying to do and using Subversion is getting in my way so I'd just like to be done with moving the core over to Git. Pure selfish motivation. Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - What matters is not ideas, but the people who have them. Good people can fix bad ideas, but good ideas can't save bad people. -- Paul Graham - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Converting Maven Core to Git
So what exactly is left to be done and can they be done now. I'm trying to finish my work and I would just not do this with subversion at all. On Oct 10, 2012, at 12:43 PM, Kristian Rosenvold kristian.rosenv...@zenior.no wrote: The repo is already ported, it's just a matter having infra flip some switches. ISTR there are some issues that need to be fixed in the core it's when running from a git repo Those issues can be fixed now ;) Kristian Den 10. okt. 2012 kl. 18:28 skrev Jason van Zyl ja...@tesla.io: Hi, Where are the scripts that were used to convert the other repos to Git? I'm going to spend tomorrow trying to convert Maven core over to Git as trying to finish the other work I'm trying to do and using Subversion is getting in my way so I'd just like to be done with moving the core over to Git. Pure selfish motivation. Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - What matters is not ideas, but the people who have them. Good people can fix bad ideas, but good ideas can't save bad people. -- Paul Graham - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - We know what we are, but know not what we may be. -- Shakespeare
Flipping Maven Core to Git
So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis
Re: Flipping Maven Core to Git
I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
I like the sound of 3.1:-) /Anders On Wed, Oct 10, 2012 at 8:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org
Re: Flipping Maven Core to Git
I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition
Re: Flipping Maven Core to Git
So you're going to be working on the core without touching the it's? Seriously, this discussion is going nowhere. We move both and someone speeds 45 minutes fixing the it's. Kristian Den 11. okt. 2012 kl. 00:54 skrev Jason van Zyl ja...@tesla.io: I think we can flip the core, we can still release the core without flipping the integration tests. It might actually be better to not move both of those at the same time. On Oct 10, 2012, at 2:36 PM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: It's just two simple projects that have some very simple fixes that need to be done (git does not preserve empty directories and I seem to recall that a couple of it's failed due to this). I say we do the 3.1 release off git ;) Move both. Kristian 2012/10/10 Anders Hammar and...@hammar.net: I see no reason they need to be migrate at the same time. /Anders On Wed, Oct 10, 2012 at 7:53 PM, Jason van Zyl ja...@tesla.io wrote: So in chatting with Kristian it appears the the core is ready to flip, while the integration tests need some work. Can we flip the core and work on the core integrations test afterward? Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - A language that doesn’t affect the way you think about programming is not worth knowing. -- Alan Perlis - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org Thanks, Jason -- Jason van Zyl Founder CTO, Sonatype Founder, Apache Maven http://twitter.com/jvanzyl - Our achievements speak for themselves. What we have to keep track of are our failures, discouragements and doubts. We tend to forget the past difficulties, the many false starts, and the painful groping. We see our past achievements as the end result of a clean forward thrust, and our present difficulties as signs of decline and decay. -- Eric Hoffer, Reflections on the Human Condition - To unsubscribe, e-mail: dev-unsubscr...@maven.apache.org For additional commands, e-mail: dev-h...@maven.apache.org