16.02.2012 21:22, KP Kirchdoerfer пишет: > Am 16.02.2012 20:00, schrieb Andrew: >> 16.02.2012 20:46, KP Kirchdoerfer пишет: >>> Am 16.02.2012 19:00, schrieb Andrew: >>>> 16.02.2012 19:54, KP Kirchdoerfer пишет: >>>>> Hi Andrew; hi all >>>>> >>>>> just to understand how we use git. >>>>> >>>>> I saw that Andrew merged next with next-experimental. >>>>> So will there be two "next" branches in the near future, or will >>>>> one(next-experimental) replace the other(next)? And if so shouldn't we >>>>> remove the old one, just one get not confused? >>>>> >>>>> kp >>>>> >>>> Merge operation is just to add all commits from one branch (or from >>>> branch till some tag/commit) to another branch. Branch structure in that >>>> case isn't modified, this is just adding changes from one branch to >>>> another. >>> Ok, but how do we decide which one is used as "main" branch for >>> developing the next version? >> 'Next' of course. Today or tomorrow I'll finish all changes (now I >> working on ipset) and merge all into 'next'. Maybe I'll update kernel too. > But you changed next into next-experimental today? And worked on > next-experimental, as I did. So you will merge next-experimental back > into "next" today or tomorrow? > And use the work of next-experimental as base to update kernel from 3.24 > to 3.2.6?? > I will not change branches, just take all commits from next-experimental to next. >>> Or is it just my decision to remove remote tracking of branch "next" >>> locally, if _I decide, that next-experimental_ is the way I'd like to go >>> with? >> Right. > IMHO "Wrong" - if I do no locally track "next" any longer, I'll be > working on the wrong branch IMHO. You in any case can add 'next' locally. You can do anything with local repo/repos till it doesn't hurt remote branches. >> I think that 'next' branch now may have 'alpha' status - enough >> stable for building and testing, and main development will be into it; >> but experimental features that may break building of 'next' branch and >> that requires cooperative work of some peoples should be in >> 'next-experimental'. Personal experiments may be done in local branches, >> that are merged with 'next'/'next-experimental', and changes from them >> may be merged completely or partial (via git cherrypick for ex.) to >> 'next'/'next-experimental' > ok. > >>> Building "new branches is cheap", tracking remote branches is easy - >>> maybe I'm just too afraid about merging? >> Maybe. GIT is versioning system that is based on patch mechanism, so >> there is no difference in resource consumption between usual commit and >> commit to new branch; also unused branches can be easily destroyed - all >> commits that weren't merged in otner branches in that case will be lost, >> but merged commits were remain on corresponding places. > I see I'll have to read some more about git and esp merging :) > > kp > Git is enough simple and very powerful system. It may look like too difficult - but it isn't too hard to understand. Look at commit like a patch, and on branch like a sequence of patches - it'll be easier to understand git in that way.
------------------------------------------------------------------------------ Virtualization & Cloud Management Using Capacity Planning Cloud computing makes use of virtualization - but cloud computing also focuses on allowing computing to be delivered as a service. http://www.accelacomm.com/jaw/sfnl/114/51521223/ _______________________________________________ leaf-devel mailing list leaf-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/leaf-devel
