OGG… I accidentally committed something to master that should have been on a brand new branch!

git commit
Author

Ryo Nakagami

Published

2024-07-12

Modified

2024-07-12

This post is based on Dangit, I accidentally committed something to master that should have been on a brand new branch!

Problem

Let’s say you have just set up a new git repository. Initially, you planned to create a new branch named “test” to work on after the first commit, which involved editing the README.md file. However, in practice, you continued to make changes on the main branch instead.

▶  What I originally planned to do

%%{init: { 'logLevel': 'debug', 'theme': 'base', 'gitGraph': {'rotateCommitLabel': true}} }%%
    gitGraph
       commit id: "813faa2"
       branch test
       commit id: "28a8bf7"
       commit id: "466fbe7"
       commit id: "30bf56a"
       commit id: "6159140"

▶  What I’ve actually done…

%%{init: { 'logLevel': 'debug', 'theme': 'base', 'gitGraph': {'rotateCommitLabel': true}} }%%
    gitGraph
       commit id: "813faa2"
       commit id: "28a8bf7"
       commit id: "466fbe7"
       commit id: "30bf56a"
       commit id: "6159140"

Solution: using git reset

▶  Steps

  1. Create a new test branch based on the current HEAD state.
  2. Reset the main branch to the state immediately after the first commit.
  3. Switch to the test branch.

▶  Commands

## Step 1
% git branch test

## Step 2
% git reset 813faa2 --hard

## Step 3
% git switch test

▶  Explains

  • You can create a new branch based on the current HEAD without switching using git branch <branch-name>.
  • Using git reset --hard instead of git reset --soft ensures that by default, changes are not left in the current branch’s index or files.
  • If changes remain in the current branch, you would need to execute git stash or git clean.

!!WARNING

If you have already pushed to the remote repository with the wrong commits, it’s better not to execute the above solution because it could inconvenience other developers.

References