October 2014 - petsc-dev - mailman3.cels.anl.gov

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Nystrom, William David 02 Oct '14

02 Oct '14

On my local machine I have my master, bare git repo with the following config file which Jed suggested a year or so ago: % cat config [core] repositoryformatversion = 0 filemode = true bare = true sharedrepository = 1 [receive] denyNonFastforwards = true [remote "origin"] url = https://bitbucket.org/petsc/petsc fetch = +refs/heads/maint:refs/heads/maint fetch = +refs/heads/master:refs/heads/master fetch = +refs/heads/next:refs/heads/next [remote "artab"] url = ssh://ml-fey.lanl.gov/usr/projects/artab/users/wdn/repos/git/petsc fetch = +refs/*:refs/* mirror = true And for my workflow, I stay in sync with the official PETSc git repo by periodically executing the following command: git remote -v update origin And then I have various git clones from my master git repo where I do actual work. Do I need to do anything special now that the "next" branch has been unwound with either my master git repo or my git clones? In my master git repo, I have 3 additional branches - wdn_mods_next, wdn_mods_master and wdn_mods_maint where I keep any of my own modifications to those respective branches. And I keep wdn_mods_next in sync with next by periodically executing the following commands in a git clone with the wdn_mods_next branch checked out: git pull origin next resolve any conflicts and commit git push origin wdn_mods_next:wdn_mods_next So, if I do the following set of steps, should I expect them to work now that next has been unwound? 1. cd to my master bare petsc repo git remote -v update origin 2. cd to my git clone of petsc with wdn_mods_next branch checked out git pull origin next resolve any conflicts and commit git push origin wdn_mods_next:wdn_mods_next Thanks, Dave -- Dave Nystrom LANL HPC-5 Phone: 505-667-7913 Email: wdn(a)lanl.gov Smail: Mail Stop B272 Group HPC-5 Los Alamos National Laboratory Los Alamos, NM 87545 ________________________________________ From: petsc-dev-bounces(a)mcs.anl.gov [petsc-dev-bounces(a)mcs.anl.gov] on behalf of Satish Balay [balay(a)mcs.anl.gov] Sent: Thursday, October 02, 2014 7:38 AM To: petsc-dev(a)mcs.anl.gov; petsc-users(a)mcs.anl.gov Subject: Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound! On Thu, 2 Oct 2014, Satish Balay wrote: > petsc git users, > > If you are using next branch from git repository for integration or > testing or working with it in some form - i.e if you have 'next' > listed by: > > git branch > > Please delete and recreate it. i.e: > > git checkout master > git branch -D next > git checkout next > > From my check - the following feature branches are currently in next > (not yet merged to master). > > Feature branch owners, please remerge your feature branches into to > next - as you see fit. BTW: Some of you might want to rebase/cleanup some of the feature branches before merging to new next. Also some of these feature branches are from pre-petsc-3.5 timeframe. So a rebase to latest master might be something to consider. [if they are still in line for a merge to master] Satish > > $ comm -12 <(git branch -r --merged next | sort) <(git branch -r --no-merged master | sort) > origin/barry/cleanup-blaslapack > origin/barry/remove-dreal > origin/barry/remove-sidl > origin/barry/reuse-hwloc > origin/dmeiser/pullrequest-fix-cusp-bjacobi2 > origin/dmeiser/pullrequest-remove-thrust-py > origin/jed/sr-driver4 > origin/jed/vecghostgetvalues > origin/karlrupp/fix-viennacl-veccopy > origin/karpeev/fix-ksp-pcgasm > origin/knepley/feature-optcontrol-mg > origin/knepley/feature-plex-generator-args > origin/knepley/feature-plex-nasm > origin/knepley/fix-configure-petsc-dep > origin/knepley/fix-quadrature-order > origin/madams/sr-driver4 > origin/maint > origin/mlange05/fix-gmsh-coordinates > origin/next > origin/paulmullowney/aijcusp-ellpack-fix > origin/prbrune/sf-examplefix > origin/prbrune/snes-ex19vtkoutput > origin/sarich/fix-tao-jbearing-bqpip > origin/shri/ts-dae-semi-explicit > origin/shri/ts-is-for-differential-variables > origin/shri/ts-powergrid > origin/stefano_zampini/matis-preallocation > origin/stefano_zampini/pcbddc-changeofbasis > origin/stefano_zampini/pcbddc_deluxe > origin/tisaac/feature-fe-geom-dim > origin/tisaac/plex-flexible-projection > > [obviously remove maint, next from the above list] > > The previous 'next' branch is currently available as 'next-oct-2014' - > if you need it for any checks/debugging. > > Satish > > > >

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Barry Smith 02 Oct '14

02 Oct '14

It is not jumbled output. It is the output I got On Oct 2, 2014, at 11:47 AM, Jed Brown <jed(a)jedbrown.org> wrote: > Barry Smith <bsmith(a)mcs.anl.gov> writes: > >> Done, but note in my previous email with a full cut and paste of what I did I never got any messages about misbehavior > > The following looks to me like jumbled output. > > | ~/Src/petsc next $ git pull > | From bitbucket.org:petsc/petsc > | + 5fad734...6e2aac5 next -> origin/next (forced update) > | * [new branch] next-oct-2014 -> origin/next-oct-2014 > | Already up-to-date! > | Merge made by the 'recursive' strategy. > | ~/Src/petsc next $ > > It doesn't make sense to say "Already up-to-date" and that it made a merge.

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Jed Brown 02 Oct '14

02 Oct '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > Done, but note in my previous email with a full cut and paste of what I did I never got any messages about misbehavior The following looks to me like jumbled output. | ~/Src/petsc next $ git pull | From bitbucket.org:petsc/petsc | + 5fad734...6e2aac5 next -> origin/next (forced update) | * [new branch] next-oct-2014 -> origin/next-oct-2014 | Already up-to-date! | Merge made by the 'recursive' strategy. | ~/Src/petsc next $ It doesn't make sense to say "Already up-to-date" and that it made a merge.

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Jed Brown 02 Oct '14

02 Oct '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > So should I redo the process Satish said or is my current next ok? (Seems hard to believe) Redo it. And set pull.ff=only.

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Barry Smith 02 Oct '14

02 Oct '14

So should I redo the process Satish said or is my current next ok? (Seems hard to believe) On Oct 2, 2014, at 11:34 AM, Jed Brown <jed(a)jedbrown.org> wrote: > Barry Smith <bsmith(a)mcs.anl.gov> writes: > >> Too late. I likely screwed up the new next. > > fixed now

1 0

Re: [petsc-dev] does next model mess up our histories
by Barry Smith 02 Oct '14

02 Oct '14

On Oct 2, 2014, at 11:32 AM, Jed Brown <jed(a)jedbrown.org> wrote: > Barry Smith <bsmith(a)mcs.anl.gov> writes: >> I disagree. We should start a CI tester now and then work on that stuff. > > It's hardly useful without the test improvements and doing CI now just > means we hack around problems that need to be fixed properly. I disagree, there is no hacking that needs to be done that has to be fixed later. We are just launching configure, make, make test that won’t change Barry

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Barry Smith 02 Oct '14

02 Oct '14

Done, but note in my previous email with a full cut and paste of what I did I never got any messages about misbehavior On Oct 2, 2014, at 11:30 AM, Jed Brown <jed(a)jedbrown.org> wrote: > Satish Balay <balay(a)mcs.anl.gov> writes: > >> balay@asterix /home/balay/petsc (master=) >> $ git co next >> Switched to branch 'next' >> Your branch is behind 'origin/next' by 1963 commits, and can be fast-forwarded. >> (use "git pull" to update your local branch) >> balay@asterix /home/balay/petsc (next<) >> >> >> Did old next get remerged into new next? Should I rewind it again? > > https://bitbucket.org/petsc/petsc/commits/8c34c3ad2d52157761586f2bb6d4466ae… > > Barry should > > $ git config pull.ff only > > so he can stop making undesirable merges from upstream.

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Satish Balay 02 Oct '14

02 Oct '14

On Thu, 2 Oct 2014, Jed Brown wrote: > Barry Smith <bsmith(a)mcs.anl.gov> writes: > > > Too late. I likely screwed up the new next. > > fixed now And everyone who had the git repo should do the following (again - if its done already) git checkout master git branch -D next git fetch git checkout next Satish

1 0

Re: [petsc-dev] [petsc-users] petsc git next branch *is* unwound!
by Jed Brown 02 Oct '14

02 Oct '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > Too late. I likely screwed up the new next. fixed now

1 0

Re: [petsc-dev] does next model mess up our histories
by Jed Brown 02 Oct '14

02 Oct '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > I disagree. We should start a CI tester now and then work on that stuff. It's hardly useful without the test improvements and doing CI now just means we hack around problems that need to be fixed properly.

1 0