Fork me on GitHub

Git introduction: Interrupted work

Overview

Teaching: 5 min
Exercises: 10 min
Questions
  • How can Git help us to deal with interrupted work and context switching?
Objectives
  • Learn to switch context or abort work without panicking.

Frequent situation: interrupted work

We all wish that we could write beautiful perfect code. But the real world is much more chaotic:

  • You are in the middle of a Jackson-Pollock-style debugging spree with 27 modified files and debugging prints everywhere.
  • Your colleague comes in and wants you to fix/commit something right now.
  • What to do?

Git provides lots of ways to switch tasks without ruining everything.


Option 1: Stashing

The stash is the first and easiest place to temporarily “stash” things.

  • git stash will put working directory and staging area changes away. Your code will be same as last commit.
  • git stash pop will return to the state you were before. Can give it a list.
  • git stash list will list the current stashes.
  • git stash save NAME is like the first, but will give it a name. Useful if it might last a while.
  • git stash save [-p] [filename] will stash certain files files and/or by patches.
  • git stash drop will drop the most recent stash (or whichever stash you give).
  • The stashes form a stack, so you can stash several batches of modifications.

Exercise: stashes

  1. Make a change.
  2. Check status/diff, stash the change, check status/diff again.
  3. Make a separate, unrelated change which doesn’t touch the same lines. Commit this change.
  4. Pop off the stash you saved, check status/diff.
  5. Optional: Do the same but stash twice. Also check git stash list. Can you pop the stashes in the opposite order?
  6. Advanced: What happens if stashes conflict with other changes? Make a change and stash it. Modify the same line or one right above or below. Pop the stash back. Resolve the conflict. Note there is no extra commit.
  7. Advanced: what does git graph show when you have something stashed?

Option 2: Create branches

You can use branches almost like you have already been doing if you need to save some work. You need to do something else for a bit? Sounds like a good time to make a feature branch.

You basically know how to do this:

$ git checkout -b temporary  # create a branch and switch to it
$ git add <paths>            # stage changes
$ git commit                 # commit them
$ git checkout master        # back to master
                             # do your work...
$ git checkout temporary     # continue where you left off

Later you can merge it to master or rebase it on top of master and resume work.

Exercises (optional)

You already know how to do this…

  1. Optional: Go through the process above. Start a change, create new branch and store your changes. Go back to master and fix something else. Resume your work and merge the new branch.
  2. Discuss how to resume your former work. Can you git rid of a branch? Continue using it? etc.

Storing various junk you don’t need but don’t want to get rid of

It happens often that you do something and don’t need it, but you don’t want to lose it right away. You can use either of the above strategies to stash/branch it away: using branches is probably better. Note that if you try to use a branch after a long time, conflicts might get really bad but at least you have the data still.

Key points

  • There is almost never reason to clone a fresh copy to complete a task that you have in mind.