DEV Community πŸ‘©β€πŸ’»πŸ‘¨β€πŸ’»

DEV Community πŸ‘©β€πŸ’»πŸ‘¨β€πŸ’» is a community of 970,177 amazing developers

We're a place where coders share, stay up-to-date and grow their careers.

Create account Log in
Cover image for Git tip: cherry-pick with an -x
Kostas Tsalikis
Kostas Tsalikis

Posted on • Originally published at droidship.com on

Git tip: cherry-pick with an -x

TL;DR

  • cherry-pick takes a commit and applies its changes to your current branch.
  • When working with others, use -x option in order to highlight on the commit message that this is cherry-picked and not the original commit.
  • Use it sparingly, instead of a merge, when you need only one specific commit.
  • Overusing it can result in confusing history with duplicate commits, thus merge is generally preferred.

What is a cherry-pick?

When cherry-picking you basically take the commit that you’re targeting and apply its changes to your current HEAD. This will lead to a brand new commit with a different SHA and the same commit message.

git cherry-pick <target commit SHA>
Enter fullscreen mode Exit fullscreen mode

Why cherry-pick in the first place?

Let’s say that you’ve drawn yourself into a corner – you need changes from another branch, but you don’t need all the changes, just some commits. Here is the only case where cherry-pick should be applied.

Main disadvantages

Most times developers choose to merge instead of performing a cherry-pick – but why is that? Well, one must be very careful when applying cherry-pick, especially with the order that different commits are applied. When merging, you can be sure that the commits are applied with the right order, so you just have to resolve appropriately the conflicts. On the contrary, when dealing with a number of cherry-picks, you must manually find the correct order of the commits and apply them one by one according to their date.

What’s more, you end up with duplicate commits, since the name of the cherry-picked commit will be the same with the original one.

So, cherry-pick is bad because:

  1. You must manually apply the cherry-picks – depending on the number of commits that you have to apply it may be difficult to find the right order and this is surely an error prone process.
  2. Duplicate commits result in weird looking and ugly history.
  3. Let’s say someone else needs that fix in another branch, will they pick the right commit or will they cherry-pick the cherry-picked one. After all, the cherry-picked one may have a conflict that you do not need.

A contrived example of ugly history.

Below follows a contrived example of how a cherry-pick can make the history look ugly. Let’s say that you create two commits on master, A and B, and at that moment you want to create a new feature branch(named with an extra dose of imagination feature) and add a new commit C.

At the moment we have something like this.

A --- B <- (master)
          \
           \
            C <- (feature)
Enter fullscreen mode Exit fullscreen mode

Then we go back again to master and commit D. We commit E on feature, but before that, we would need to apply fix D from master, thus we cherry-pick D.

   A --- B --- D     <- (master)
          \
           \
            C --- D' --- E     <- (feature)
Enter fullscreen mode Exit fullscreen mode

At this point, we would like to merge the feature back to master. Now the graph would seem something like this:

A --- B --- D ----- Merge Commit     <- (master)
          \                /
           \              /
            C --- D' --- E     <- (feature)
Enter fullscreen mode Exit fullscreen mode

And a git log --oneline would show us the duplicate commits:

5b5f739 (HEAD -> master) Merge branch 'feature'
d5c76ad E
52ed367 D
c097b61 D
af599fa C
383a1e1 B
c1fffad A
Enter fullscreen mode Exit fullscreen mode

Of course, in this example we could use rebase in order to avoid duplicate commits, but one can think of a scenario where the second branch would be a long lived one and would need for only a hotfix to be applied from the upstream.

Working with others

So, what do you do in case that you are not the only one that needs these precious fixes? Naturally, the other person will try to search, with the commit message, the desired commit. Guess what, you will find more than one commit and you won’t know which one to apply. One could of course try to change the commit name so it can hint that it’s a cherry-pick commit and not the original one. This can be done with the following command:

git cherry-pick <target commit SHA> -e
Enter fullscreen mode Exit fullscreen mode

This will open an editor where you can change the commit message and append something that hints that this is a cherry-pick commit.

After some digging I found that there is another option which automatically appends (cherry picked from commit <original commit's SHA>)

git cherry-pick <target commit SHA> -x
Enter fullscreen mode Exit fullscreen mode

At least now you know which commit is cherry-picked or not and what was the original commit as well.

Cover image by Robert Zunikoff on Unsplash

Top comments (0)

🌚 Friends don't let friends browse without dark mode.

Sorry, it's true.