[Bioc-devel] GitHub and Bioc sync: issue with dup commit for the sva package

Thu Aug 30 00:03:22 CEST 2018

Thank you, Leo!

On Mon, Aug 27, 2018 at 1:53 PM Leonardo Collado Torres <lcollado using jhu.edu>
wrote:

> Hi Nitesh,
>
> After looking more closely, I went with option 1 that you clearly
> outlined below. The cherry picked version of the history is now the
> origin/master one at
> https://github.com/jtleek/sva-devel/commits/master. From
> http://bioconductor.org/developers/how-to/git/resolve-duplicate-commits/
> my understanding is that now you have to force the Bioconductor git
> history (upstream/master with the default setup) to match the one from
> the sva-devel GitHub repo.
>
> The only complicated commit was
>
> https://github.com/jtleek/sva-devel/commit/e6b80f62d494f77465650cf3d624feaf8913e837
> where I had to use the "theirs" strategy for many files.
>
> At the end I added this commit
>
> https://github.com/jtleek/sva-devel/commit/05cfab55c893b4280aaaa61ab703f3abd235889a
> that bumps the version of the package to 3.29.1 on bioc-devel to
> propagate Claire's recent changes.
>
> Best,
> Leo
>
> PS I tried the git push -f upstream/master out of curiousity, and yup,
> I can't do it:
>
> $ git push -f upstream master
> Counting objects: 111, done.
> Delta compression using up to 4 threads.
> Compressing objects: 100% (54/54), done.
> Writing objects: 100% (111/111), 19.58 KiB | 2.80 MiB/s, done.
> Total 111 (delta 83), reused 62 (delta 56)
> remote: Resolving deltas: 100% (83/83), completed with 38 local objects.
> remote: FATAL: + refs/heads/master packages/sva l.collado-torres
> DENIED by fallthru
> remote: error: hook declined to update refs/heads/master
> To git.bioconductor.org:packages/sva.git
>  ! [remote rejected] master -> master (hook declined)
> error: failed to push some refs to 'git using git.bioconductor.org:
> packages/sva.git'
>
> PS2 I did push the master_backup branch to GitHub just in case we need
> it later https://github.com/jtleek/sva-devel/tree/master_backup
>
>
> On Sat, Aug 25, 2018 at 2:00 AM Turaga, Nitesh
> <Nitesh.Turaga using roswellpark.org> wrote:
> >
> > Hi Leo,
> >
> > I’ll try to help you with this issue.
> >
> > The duplicates commit you have on your package “sva” come from the
> pre-git era of Bioconductor.  There are three ways we can fix this,
> >
> > 1,
> >
> > You fix your GitHub repo by removing the duplicate commits. They happen
> only after this commit.
> >
> > commit d12b53824915f0ff7b7906043c7fc9237521e8f3
> > Author: Kipper Fletez-Brant <cafletezbrant using gmail.com>
> > Date:   Sat Apr 22 10:42:29 2017 -0400
> >
> >     actually did it
> >
> >
> > You can do the cherry-picking of the commits (only one from a set of
> duplicate commits, one after the other to construct your git history). It
> is outlined in this document here,
> >
> >
> http://bioconductor.org/developers/how-to/git/resolve-duplicate-commits/.
> >
> > If you choose this method, I can help with any questions you may have.
> >
> >         Some help if you decide to choose this route,
> >
> >         A.
> >
> >         You might face issues with this commit (7e4aa44 Merge pull
> request #26 from cafletezbrant/master) while doing the cherry-pick.
> >         You can ignore this, because the commits themselves will be in
> the git history.
> >
> >         d12b538 actually did it
> >         7cedd49 needed to use tcrossprod(t(...)) a bunch
> >
> >
> >         B.
> >
> >         If you hit a lot of merge conflicts while doing the cherry-pick,
> the best way I’ve gotten used to is by using
> >
> >                 `git mergetool`.
> >
> >         (I personally use the option `git mergetool -t vimdiff`, but
> feel free to use whatever merge tool you’d like)
> >
> >         C.
> >
> >         Make sure you cherry-pick the RELEASE commits which come from
> the core team while doing the cherry-pick process.
> >
> > 2,
> >
> > I can fix the bioconductor repo on my end to show no duplicate commits.
> Then you’d have to force your Github master branch to be the same as the
> Bioconductor repo and then put Claire’s commits on top of them. This is a
> slightly less optimal solution because I might make some mistakes resolving
> the conflicts, since I don’t know the details of code.
> >
> > 3,
> >
> > Since the issue of duplicate commits was pre-git era, we can just ignore
> the duplicates. It is pretty straightforward to disable the duplicate
> commit check, but it should only be a worst case. I don’t recommend this
> for your repo because there are very few issues and it seems we can resolve
> them. I reserve this option for packages which have extensively fractured
> their git history.
> >
> >
> > Best regards,
> >
> > Nitesh
> >
> >
> > =
> > Nitesh Turaga
> > Bioconductor Core Team
> >
> >
> > > On Aug 24, 2018, at 11:29 PM, Leonardo Collado Torres <
> lcollado using jhu.edu> wrote:
> > >
> > > Hi,
> > >
> > > I'm having trouble with synching a github repo with Bioc.
> > >
> > > At https://github.com/jtleek/sva-devel/commits/master we have 3 recent
> > > commits by Claire that we want to sync with Bioc. I followed the
> > > instructions at
> > >
> http://bioconductor.org/developers/how-to/git/sync-existing-repositories/
> > > and then had to do step #8 where I ran steps 1-4 from "force
> > > Bioconductor master to Github master"
> > >
> http://bioconductor.org/developers/how-to/git/abandon-changes/#force-bioconductor--to-github-
> .
> > > On the optional step 4 I used git cherry-pick to pick the 3 commits by
> > > Claire. That's how we got to the current state on the jtleek/sva-devel
> > > repo (git added me as a co-author of the commits). However, "git push
> > > upstream master" keeps failing as shown below:
> > >
> > > $ git push upstream master
> > > Counting objects: 15, done.
> > > Delta compression using up to 4 threads.
> > > Compressing objects: 100% (7/7), done.
> > > Writing objects: 100% (15/15), 2.76 KiB | 2.76 MiB/s, done.
> > > Total 15 (delta 9), reused 11 (delta 8)
> > > remote: Error: duplicate commits.
> > > remote:
> > > remote: There are duplicate commits in your commit history, These
> cannot be
> > > remote: pushed to the Bioconductor git server. Please make sure that
> this is
> > > remote: resolved.
> > > remote:
> > > remote: Take a look at the documentation to fix this,
> > > remote:
> https://bioconductor.org/developers/how-to/git/sync-existing-repositories/
> ,
> > > remote: particularly, point #8 (force Bioconductor master to Github
> master).
> > > remote:
> > > remote: For more information, or help resolving this issue, contact
> > > remote: <bioc-devel using r-project.org>. Provide the error, the package
> name and
> > > remote: any other details we might need.
> > > remote:
> > > remote: Use
> > > remote:
> > > remote:     git show 812b53941d5ba3e70797e4ecfef8401b89fc4741
> > > remote:     git show a3406ad11056cc5f1896f71c7864e2c3bd80b817
> > > remote:
> > > remote: to see body of commits.
> > > remote:
> > > To git.bioconductor.org:packages/sva.git
> > > ! [remote rejected] master -> master (pre-receive hook declined)
> > > error: failed to push some refs to 'git using git.bioconductor.org:
> packages/sva.git'
> > >
> > >
> > > If I clone git using git.bioconductor.org:packages/sva.git I can see that
> > > both 812b5394 and a3406ad1 already exist on the Bioc git. I see that
> > > one of the notes at
> > >
> http://bioconductor.org/developers/how-to/git/abandon-changes/#force-bioconductor--to-github-
> > > mentions that the step 1-4 solution won't work in this scenario. The
> > > "Reset to a previous commit" section recommends using git reset
> > > --hard. But that still fails.
> > >
> > > $ git reset -- hard a3406ad11056c
> > > 13:58 sva-devel $ git status
> > > On branch master
> > > Your branch is ahead of 'upstream/master' by 3 commits.
> > >  (use "git push" to publish your local commits)
> > >
> > > nothing to commit, working tree clean
> > > 13:58 sva-devel $ git push -f upstream master
> > > Counting objects: 15, done.
> > > Delta compression using up to 4 threads.
> > > Compressing objects: 100% (7/7), done.
> > > Writing objects: 100% (15/15), 2.76 KiB | 1.38 MiB/s, done.
> > > Total 15 (delta 9), reused 11 (delta 8)
> > > remote: Error: duplicate commits.
> > > remote:
> > > remote: There are duplicate commits in your commit history, These
> cannot be
> > > remote: pushed to the Bioconductor git server. Please make sure that
> this is
> > > remote: resolved.
> > > remote:
> > > remote: Take a look at the documentation to fix this,
> > > remote:
> https://bioconductor.org/developers/how-to/git/sync-existing-repositories/
> ,
> > > remote: particularly, point #8 (force Bioconductor master to Github
> master).
> > > remote:
> > > remote: For more information, or help resolving this issue, contact
> > > remote: <bioc-devel using r-project.org>. Provide the error, the package
> name and
> > > remote: any other details we might need.
> > > remote:
> > > remote: Use
> > > remote:
> > > remote:     git show 812b53941d5ba3e70797e4ecfef8401b89fc4741
> > > remote:     git show a3406ad11056cc5f1896f71c7864e2c3bd80b817
> > > remote:
> > > remote: to see body of commits.
> > > remote:
> > > To git.bioconductor.org:packages/sva.git
> > > ! [remote rejected] master -> master (pre-receive hook declined)
> > > error: failed to push some refs to 'git using git.bioconductor.org:
> packages/sva.git'
> > >
> > > So hm... I have no idea how to fix this.
> > >
> > > Best,
> > > Leo
> > >
> > > _______________________________________________
> > > Bioc-devel using r-project.org mailing list
> > > https://stat.ethz.ch/mailman/listinfo/bioc-devel
> >
> >
> >
> > This email message may contain legally privileged and/or confidential
> information.  If you are not the intended recipient(s), or the employee or
> agent responsible for the delivery of this message to the intended
> recipient(s), you are hereby notified that any disclosure, copying,
> distribution, or use of this email message is prohibited.  If you have
> received this message in error, please notify the sender immediately by
> e-mail and delete this email message from your computer. Thank you.
>

	[[alternative HTML version deleted]]