DEV Community

loading...

Git quick tips #2: Working with many branches

Michael Kohl
I dev @ DEV. Your friendly neighborhood anarcho-cynicalist. ¯\_(ツ)_/¯ and (╯°□°)╯︵ ┻━┻) are my two natural states.Tag mod for #ruby, #fsharp, #ocaml
Originally published at citizen428.net ・2 min read

I somehow often end up working on a few branches in parallel and some things make my life easier:

  1. git switch -: Like the cd command git switch and git checkout take a - argument which takes you back to the previously checked out branch.

    $ git branch
    * citizen428/long-branch-name
    $ git switch main
    Switched to branch 'main'
    Your branch is up to date with 'origin/main'.
    $ git switch -
    Switched to branch 'citizen428/long-branch-name'
    Your branch is up to date with 'origin/citizen428/long-branch-name'.
    

    I actually have checkout aliased to co and switch to sw (as well as git itselftog`) but I didn't want to confuse readers with those.

  2. git show branch:file_name: Especially during some of my recent work I often had several branches that modified the same files. A quick way to see a file from another branch without having to check it out is git show, e.g. git show main:app/models/user.rb.

  3. git-rerere: Short for "reuse recorded resolution". Particularly useful for long-running branches (which ideally we shouldn't have but ideals don't tend to survive contact with reality). The process is automatic and doesn't generally require manual intervention, just make sure it's enabled in your git configuration, e.g. git config --global rerere.enabled true. The git docs have a very nice explanation.

  4. git stash branch: I sometimes start working in the wrong checkout and a quick and easy way to correct this is the following:

    $ git stash # stash current changes
    $ git stash branch  
    

    This will directly create a new branch from the stash instead of having to first create it and then using git stash pop. Keep in mind that the currently active branch will be used as the base, so unless you're already on main you may want to check that out first.

  5. git autocorrect: Not directly related to branch management but it can be useful for the commands I use rarely and have not aliased. You can configure it like this git config --global help.autocorrect 10, where the integer is a value in 10th of a second, so the above will wait for 1 second before running the autocorrected command:
    autocorrect in action

Discussion (5)

Collapse
ardunster profile image
Anna R Dunster

Didn't know about that autocorrect feature, that sounds handy.

Collapse
matthewpersico profile image
Matthew O. Persico

Disk is cheap. I put every new branch in its own same-named worktree. Switching tasks is as easy as a 'cd' command. No stashing, no worrying about what branch am I on because its in my pwd which is in my prompt.

Collapse
citizen428 profile image
Michael Kohl Author

That sounds like a flashback to SVN branches. 🙂 But why not, if it works for you it's a good system.

Collapse
matthewpersico profile image
Matthew O. Persico

It’s very highly scripted. I keep all gits in $HOME/ gits/ “remote server”/org/repo. Each worktree in a subdir of that called wt. I can run a command git go to give me a listing and then I pick one to cd to. I even wrote my own ‘git’ dispatcher function so I could enhance existing commands. I should write an article about that

Thread Thread
citizen428 profile image
Michael Kohl Author

I for one would love to read it! IMO one of Git's best features is what you can build on top of it in terms of highly individualized workflows, it's almost like a VCS construction kit.