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

Re: [petsc-dev] bitbucket giving wrong information
by Barry Smith 04 Dec '14

04 Dec '14

> On Dec 4, 2014, at 5:08 PM, Jed Brown <jed(a)jedbrown.org> wrote: > > Barry Smith <bsmith(a)mcs.anl.gov> writes: >> Sure and the first step of the workflow is knowing how to create >> that local branch. So how were these created (exactly) and where >> is that automated/documented? > > I do > > git fetch bitbucket:username/petsc their-branch-name:username/my-branch-name-for-their-branch > git push -u origin username/my-branch-name-for-their-branch > If we are doing this to generate a local branch that people can work with then we need to be more systemic with no my-branch-name stuff. If follow Dominics approach we should do > git fetch bitbucket:username/petsc their-branch-name:username/pullrequest-their-branch-name > git push -u origin username/pullrequest-their-branch-name If this works then lets document it in the wiki and be done with it. Satish, could you check this and document in the wiki. (it goes with our discussion of handling pull requests). Thanks Barry

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Jed Brown 04 Dec '14

04 Dec '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > Sure and the first step of the workflow is knowing how to create > that local branch. So how were these created (exactly) and where > is that automated/documented? I do git fetch bitbucket:username/petsc their-branch-name:username/my-branch-name-for-their-branch git push -u origin username/my-branch-name-for-their-branch

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Barry Smith 04 Dec '14

04 Dec '14

> On Dec 4, 2014, at 4:10 PM, Satish Balay <balay(a)mcs.anl.gov> wrote: > > On Thu, 4 Dec 2014, Barry Smith wrote: > >> >>> On Dec 4, 2014, at 3:33 PM, Jed Brown <jed(a)jedbrown.org> wrote: >>> >>> Patrick Sanan <patrick.sanan(a)gmail.com> writes: >>>>> Sure thanks, but there is no way I can remember this; hence bitbucket should walk me through it. >>>> Yeah - their helpful hint doesnt work when pulling across repos, which is a useful, probably typical, case, so it'd be better for then to have said nothing! >>> >>> GitHub does better here: >>> >>> | You can merge this Pull Request by running: >>> | >>> | git pull https://github.com/someuser/petsc someuser/branch-name >> >> Nice, this is what I want. Has bitbucket been made aware of their definiciecy so they can fix it? > > I think we need a workflow where we create a localbranch for each of > these pull requests - not directly merge (in case we need to make > local fixes). Sure and the first step of the workflow is knowing how to create that local branch. So how were these created (exactly) and where is that automated/documented? Barry > > I see some such branches currently used.. > > $ git branch -r |grep pull > origin/dmeiser/pullrequest-fix-PetscOptionsEnum > origin/dmeiser/pullrequest-fix-PetscOptionsEnum2 > origin/dmeiser/pullrequest-fix-cusp-bjacobi2 > origin/dmeiser/pullrequest-fix-dmda-cusparse > origin/dmeiser/pullrequest-fix-format-specifier-ex43 > origin/dmeiser/pullrequest-remove-thrust-py > > > Satish

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Satish Balay 04 Dec '14

04 Dec '14

On Thu, 4 Dec 2014, Barry Smith wrote: > > > On Dec 4, 2014, at 3:33 PM, Jed Brown <jed(a)jedbrown.org> wrote: > > > > Patrick Sanan <patrick.sanan(a)gmail.com> writes: > >>> Sure thanks, but there is no way I can remember this; hence bitbucket should walk me through it. > >> Yeah - their helpful hint doesnt work when pulling across repos, which is a useful, probably typical, case, so it'd be better for then to have said nothing! > > > > GitHub does better here: > > > > | You can merge this Pull Request by running: > > | > > | git pull https://github.com/someuser/petsc someuser/branch-name > > Nice, this is what I want. Has bitbucket been made aware of their definiciecy so they can fix it? I think we need a workflow where we create a localbranch for each of these pull requests - not directly merge (in case we need to make local fixes). I see some such branches currently used.. $ git branch -r |grep pull origin/dmeiser/pullrequest-fix-PetscOptionsEnum origin/dmeiser/pullrequest-fix-PetscOptionsEnum2 origin/dmeiser/pullrequest-fix-cusp-bjacobi2 origin/dmeiser/pullrequest-fix-dmda-cusparse origin/dmeiser/pullrequest-fix-format-specifier-ex43 origin/dmeiser/pullrequest-remove-thrust-py Satish

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Jed Brown 04 Dec '14

04 Dec '14

Barry Smith <bsmith(a)mcs.anl.gov> writes: > Nice, this is what I want. Has bitbucket been made aware of their definiciecy so they can fix it? I don't know, but it seems to me that they consciously avoid putting useful information in notification emails. For example, PRs between branches in the same repo don't even list commits (branches from a different repo have a short list) and none have a diffstat. "git request-pull" contains this information (as does GitHub), so they are either ignorant or they decided not to include it.

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Barry Smith 04 Dec '14

04 Dec '14

> On Dec 4, 2014, at 3:33 PM, Jed Brown <jed(a)jedbrown.org> wrote: > > Patrick Sanan <patrick.sanan(a)gmail.com> writes: >>> Sure thanks, but there is no way I can remember this; hence bitbucket should walk me through it. >> Yeah - their helpful hint doesnt work when pulling across repos, which is a useful, probably typical, case, so it'd be better for then to have said nothing! > > GitHub does better here: > > | You can merge this Pull Request by running: > | > | git pull https://github.com/someuser/petsc someuser/branch-name Nice, this is what I want. Has bitbucket been made aware of their definiciecy so they can fix it? Barry

1 0

Re: [petsc-dev] DMGetDS in fortran
by Matthew Knepley 04 Dec '14

04 Dec '14

On Sun, Nov 30, 2014 at 4:40 PM, Adrian Croucher <a.croucher(a)auckland.ac.nz> wrote: > hi > > On 29/11/14 01:43, Matthew Knepley wrote: > > Every time we add a type to PETSc, we have to tell the Fortran stub > generator about it. This is a clunky part of the > design, but there is not enough momentum to change it. Can you try > deleting your PETSC_ARCH directory and > reconfiguring? FIRST, > > cd $PETSC_DIR > mv $PETSC_ARCH/conf/reconfigure-$PETSC_ARCH.py . > > so that you can get the same configuration. > > > Thanks, that has done the trick! I must remember that in case I run into > it again. > > > >> Also, any chance of a fortran interface to DMPlexCreateFromFile()? It >> appears to be missing at the moment. >> > > Yes, I can add this. > > Added to next. Thanks, Matt > Thanks again, much appreciated. > > Cheers, Adrian > > -- > Dr Adrian Croucher > Senior Research Fellow > Department of Engineering Science > University of Auckland, New Zealand > email: a.croucher(a)auckland.ac.nz > tel: +64 (0)9 923 84611 > > -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Jed Brown 04 Dec '14

04 Dec '14

Patrick Sanan <patrick.sanan(a)gmail.com> writes: >> Sure thanks, but there is no way I can remember this; hence bitbucket should walk me through it. > Yeah - their helpful hint doesnt work when pulling across repos, which is a useful, probably typical, case, so it'd be better for then to have said nothing! GitHub does better here: | You can merge this Pull Request by running: | | git pull https://github.com/someuser/petsc someuser/branch-name

1 0

Re: [petsc-dev] bitbucket giving wrong information
by Patrick Sanan 04 Dec '14

04 Dec '14

> Il giorno Dec 4, 2014, alle ore 10:13 PM, Barry Smith <bsmith(a)mcs.anl.gov> ha scritto: > > >> On Dec 3, 2014, at 10:43 PM, Patrick Sanan <patrick.sanan(a)gmail.com> wrote: >> >> Whoops, guess that wasn't the cause then! >> >> Actually, looking at bitbucket, it looks like the PR is coming from a different repository (https://bitbucket.org/iaannn/petsc/) >> >> This worked on my machine: >> [petsc (master=)]$ git remote add iaannn [email protected]:iaannn/petsc.git >> [petsc (master=)]$ git fetch iaannn >> [petsc (master=)]$ git checkout -b ianwilliamson/fix-matlab-write-bin-complex iaannn/ianwilliamson/fix-matlab-write-bin-complex >> [petsc (ianwilliamson/fix-matlab-write-bin-complex=)]$ > > Sure thanks, but there is no way I can remember this; hence bitbucket should walk me through it. Yeah - their helpful hint doesnt work when pulling across repos, which is a useful, probably typical, case, so it'd be better for then to have said nothing! > > Barry > >> >>> On 04/12/14 12:57 PM, Barry Smith wrote: >>> $ git checkout -b ianwilliamson/fix-matlab-write-bin-complex origin/ianwilliamson/fix-matlab-write-bin-complex >>> fatal: Cannot update paths and switch to branch 'ianwilliamson/fix-matlab-write-bin-complex' at the same time. >>> Did you intend to checkout 'origin/ianwilliamson/fix-matlab-write-bin-complex' which can not be resolved as commit? >>> >>> >>> >>>> On Dec 3, 2014, at 9:54 PM, Patrick Sanan <patrick.sanan(a)gmail.com> wrote: >>>> >>>> Is it possible that you have a second remote with a branch with the same name? In that case, perhaps something like: >>>> >>>> git checkout -b ianwilliamson/fix-matlab-write-bin-complex origin/ianwilliamson/fix-matlab-write-bin-complex >>>> >>>>> On 04/12/14 12:37 PM, Barry Smith wrote: >>>>> So some guy posts a pull request. I, in trying to follow PETSc-rules, want to check out his branch and then merge that rather than use the merge directly from the bitbucket website. So I go to the branch in the bitbucket website and bitbucket claims I can look at it with >>>>> >>>>> git fetch && git checkout ianwilliamson/fix-matlab-write-bin-complex >>>>> >>>>> (see attachment). I try that and of course it does not work: >>>>> >>>>> $ git fetch && git checkout ianwilliamson/fix-matlab-write-bin-complex >>>>> remote: Counting objects: 56, done. >>>>> remote: Compressing objects: 100% (56/56), done. >>>>> remote: Total 56 (delta 48), reused 0 (delta 0) >>>>> Unpacking objects: 100% (56/56), done. >>>>> From bitbucket.org:petsc/petsc >>>>> f685883..cf8fc5e master -> origin/master >>>>> 7cadcfe..39ea070 knepley/fix-plex-orient -> origin/knepley/fix-plex-orient >>>>> 565f831..ce993ff knepley/pylith -> origin/knepley/pylith >>>>> + 3f11d05...21345e8 mlange/refactor-plex-distribute -> origin/mlange/refactor-plex-distribute (forced update) >>>>> e9d5acc..916966e next -> origin/next >>>>> error: pathspec 'ianwilliamson/fix-matlab-write-bin-complex' did not match any file(s) known to git. >>>>> >>>>> So is bitbucket lying to me or am I doing something wrong. Should this be a feature request to bitbucket ? Remember I am living in the world where I cannot remember how to do anything and thus expect my tools to tell me how to do something. So why cannot bitbucket tell me the truth of how to get this branch? >>>>> >>>>> Barry >>>>> >>>>> >>>>> >>>>> >>>>> <Mail Attachment.png> >

1 0

Re: [petsc-dev] [EXTERNAL] [pflotran-dev: 2718] Unintended consequence of CHKERRQ()
by Jed Brown 04 Dec '14

04 Dec '14

Gautam Bisht <gbisht(a)lbl.gov> writes: > Hi Jed, > > Earlier I completely missed your point. I believe the utility of > SNESSetFunctionDomainError > is that an application can determine if the state (xx) being passed to > compute residual function is meaningful (eg. a pair of (P,T) could result > in negative density). If the state is not meaningful, the application could > simply call SNESSetFunctionDomainError() and not update the residual > vector. This would cause SNES to exit with SNES_DIVERGED_FUNCTION_DOMAIN. Yes. > Looking at PFLOTRAN code more carefully, I now realize that Peter is > using SNESSetFunctionDomainError() > in FLASH2/MPHASE/MISCIBLE/IMMISCIBLE modes. I will now implement similar > check for TH mode. Great.

1 0