Hi Bram, Today I discussed this with Angelo and Marcel. In the Apache Felix projects first a 'release proposal' is created for a certain release. Once this proposal is approved committers are working towards this release, still committing on the trunk. In fact no branches are made at all, the trunk is just tagged as 'release candidate 1' at the moment that all issues have been closed. It is certainly not the intention that during this release preparation code is committed to the trunk that is not intended to be released (i.e. PoC code), which would disturb this process. As no branches are created at all (for overhead reasons; merging changes back and forth between branch and trunk) committing PoC code is definitely not appreciated. We agreed that we should have the same approach for Amdatu. As we are working towards the 0.0.6 release, any commits not intended to be released with 0.0.6 should be withholded. If you need a backup, commit it to your own sandbox, create a branch for it, or whatever. BTW, I promised to create the 0.0.6 branch yesterday but due to network outage and still failing integration tests (AMDATU-189) I didn't do that yet. I first want to resolve 189 and 76 before I branch, as this reduces the overhead involved in branching back and forth between branch and trunk.
Regards, Ivo -----Original Message----- From: amdatu-developers-bounces at amdatu.org [mailto:[email protected]] On Behalf Of Ivo Ladage-van Doorn Sent: dinsdag 30 november 2010 13:50 To: amdatu-developers at amdatu.org Subject: Re: [Amdatu-developers] Artifact naming conventions I agree that we should branch, and I was planning to do so. You could have made the branch yourself, and then commit the changes on the trunk instead of waiting for the branch to be made and then revert these changes in that branch. Anyway, I'll create a 0.0.6 branch today. Regards, Ivo -----Original Message----- From: amdatu-developers-bounces at amdatu.org [mailto:[email protected]] On Behalf Of Bram de Kruijff Sent: dinsdag 30 november 2010 13:26 To: amdatu-developers at amdatu.org Subject: Re: [Amdatu-developers] Artifact naming conventions On Tue, Nov 30, 2010 at 12:17 PM, Ivo Ladage-van Doorn <Ivo.Ladage-vanDoorn at gxsoftware.com> wrote: > I think it's a little bit strange that on one hand we are branching of 0.0.6, > use release candidates and even voting for the final release but on the other > hand anyone can drop any code at any time in the trunk, which may or may not > become (depending on the moment 0.0.6 is branched off) part of the 0.0.6 svn > tree. Not exactly sure what you mean. Let me try .. > I think it's a little bit strange that on one hand we are branching of 0.0.6 Are we? There is no release branch which is why we got in this situation in the first place. > use release candidates Are we? There are no RC tags for 0.0.6 (yet?) > even voting for the final release Which is good, but that is unrelated to in what branch the code is stored. > anyone can drop any code at any time in the trunk, Which is good. You do not want to be broadcasting code freezes on trunk, certainly not for an extended time. Common practice is that when you think that a version is feature complete you branch it into a release branch. In the branch you can tweak it for release (like reverting trunk commits that you do not want in the release, settings versions, last minute bug fixes etc). >From the release branch you can tag alpha, beta, RC and final release. Once we have Maven release management up and running, this is exactly what will happen. Again, you may rightfully argue that my commit should be in a sandbox because of the 'preliminary' state it is in (and I am happy to move it). However, the principle argument is that you should not be blocking the trunk, because... well it is the trunk. > I reckon that it's not in the reactor build so it can't (shouldn't) effect > any build or release. So we could agree on that this is allowed if and only > if the committed stuff is not in the reactor build. As we are bootstrapping I can live with that, but we'll get into the same discussion next time I commit a change on an artifact that is part of the release. IMHO bottomline is that we should branch just like the rest of the world. Regards, Bram _______________________________________________ Amdatu-developers mailing list Amdatu-developers at amdatu.org http://lists.amdatu.org/mailman/listinfo/amdatu-developers _______________________________________________ Amdatu-developers mailing list Amdatu-developers at amdatu.org http://lists.amdatu.org/mailman/listinfo/amdatu-developers
