[Bioc-devel] Do not push with merge conflicts to Bioconductor git repo.

Nitesh Turaga ntur@g@@b|oc @end|ng |rom gm@||@com
Wed May 13 18:27:21 CEST 2020


Dear Developers,

Every time there is a release for Bioconductor we (the core team), face an issue where developers end up pushing code with the "merge conflicts". One of the reasons why we get the issue "Bad description" file in the build report is because of a merge which has gone bad and produced conflicts. 

To avoid this issue, we have taken a step to include a pre-receive hook in git which checks for the merge conflict markers and rejects your push.

Please note that, although this will keep our end of things (the Bioconductor git repo) clean, your local branch will need to be fixed before you try and push to Bioconductor again. To avoid having your push to Bioconductor rejected, it's easy to employ some good practices,

1. Check your code and DESCRIPTION file for merge conflict markers, i.e for anything that looks like

        ++<<<<<<< HEAD 
          +Version: 1.1.2 
        ++======= 
          + Version: 1.0.1 
        ++>>>>>>> origin/RELEASE_3_11

   Notice the <<<<<<< HEAD, =======, >>>>>>> markers in the files.

2. Before a commit, you can use "git diff --check", and it will show you the issues with your code.

       git diff --check                                           
       
      DESCRIPTION:4: leftover conflict marker
      DESCRIPTION:6: leftover conflict marker
      DESCRIPTION:8: leftover conflict marker

3. If you use an IDE there are certain "extensions" you can add to your IDE to make it detect merge conflicts for you automatically. Most editors have them (emacs, vim, visual-studio, atom, sublime etc). Please use these if they are your development environments. You can even get built in ones for bash or zsh.

If you need help resolving merge conflicts, please check,

	https://help.github.com/en/github/collaborating-with-issues-and-pull-requests/addressing-merge-conflicts

This is mostly an effort to avoid packages from breaking for needless reasons. As usual, thanks to everyone for the contributions.


Best regards,

Nitesh 


More information about the Bioc-devel mailing list