git|June 11, 2020|3 min read

Git - How to create a Pull Request with no history of commits

TL;DR

Use git rebase or similar techniques to squash multiple intermediate commits into a single commit before creating a pull request, keeping PR history clean for reviewers.

Git - How to create a Pull Request with no history of commits

Introduction

If you working on a github project in a team. Consider you have created a branch from Master branch, and did few commits as we normally do.

When the feature work is done on that branch. You would want to commit that in Master branch. But, when you create a Pull Request (PR) for that merge. Reviewer will see the intermediate commits in your PR. Which might annoy him. And there is no benefit of incorporating that history into Master branch. Although, there is a “Squash and Merge” option. But, anyone can easily forget that because this option is not enabled by default.

So, reviewer would want to cleanup your commit history so that only one commit is visible in the PR.

This post is about create a PR without historical commits. So that, reviewer will only see single commit containing all the file changes.

Method-1 Using Git rebase

Apart from git merge, there is a git rebase option too. Git merge and git rebase both are used to integrate your changes into main branch.

# Switch to your feature branch
git checkout FeatureBranch

# in this case, our main branch is master
git rebase master

If there is no conflicts, this will work. Now, you just to push the changes.

git push -f

Method-2 Using Git rebase with interactive option

We will use interactive way of doing git rebase. Interactive rebasing is used to manipulate your history. And we would want to clean up history.

# Switch to your feature branch
git checkout FeatureBranch
# start interactive rebasing
# in this case, our main branch is master
git rebase -i master

This will open up an editor (vim kind) and few options. See example below:

pick 8472as3 add Test bug-1
pick 92731cf add feature description
pick asf4234 bug fix
# Commands:
# p, pick = use commit
# r, reword = use commit, but edit the commit message
# e, edit = use commit, but stop for amending
# s, squash = use commit, but meld into previous commit
# f, fixup = like "squash", but discard this commit's log message
# x, exec = run command (the rest of the line) using shell

It start with your commits list. Now you have to edit this. Note the default: pick keyword in front of each commits. And, below that there are various options. Now, we want to have all the commits into one. Put pick keyword with only one commit, that will be there in final history. Put squash keyword in-place of pick in front of commits. The file looks like:

pick dae2691 add amazing feature
squash 3491879 add test
squash 3fedbd5 fix typo
# Commands:
# p, pick = use commit
# r, reword = use commit, but edit the commit message
# e, edit = use commit, but stop for amending
# s, squash = use commit, but meld into previous commit
# f, fixup = like "squash", but discard this commit's log message
# x, exec = run command (the rest of the line) using shell

Save it, and quit editor. On console, it will show some messages. Messages ends with something like:

Successfully rebased and updated refs/heads/FeatureBranch.

Now you have to push your changes. Just run:

git push -f

Now create your PR, and you will see only one commit in history.

Bonus

If your feature branch is forked from another git repo. Use below commands:

git checkout FeatureBranch
git remote add upstream https://github.com/<main_repo>/PROJECT.git
git rebase -i upstream/<main_branch_name>
# after saving editor content.
git push -f

Note: You might see some conflicts, which you need to resolve before pushing changes.

Method-3 How to create PR without history by creating another branch

Assume, You created a branch named: FeatureBranch from Master branch. And you did some commits to it.

Git commits

Now, create another branch from Master branch. Lets name it:

feature_branch_no_history

Now, You have three branches:

  • master
  • feature_branch (your active branch, having commits)
  • feature_branch_no_history (same as Master)

Now, we will do this in two steps:

  1. Now, create a pull request for merging feature_branch into feature_branch_no_history and merge it by using “Squash and Merge” way.
  2. Create another pull request from feature_branch_no_history to master branch.

Git commits

Note: you will see more than one commits in this process too.

Happy code reviewing. Happy reviewer… Happy Coding…

Related Posts

Shell script to extract single file from a bunch of jar files

Shell script to extract single file from a bunch of jar files

Each jar file has a manifest file in META_INF/MANIFEST.MF I have to read each…

How to Solve Spring Okta/Saml issue of SAML message intended destination endpoint did not match the recipient endpoint

How to Solve Spring Okta/Saml issue of SAML message intended destination endpoint did not match the recipient endpoint

Introduction I was trying to integrate Okta with Spring, and when I deploy the…

How to Fetch Multiple Credentials and Expose them in Environment using Jenkinsfile pipeline

How to Fetch Multiple Credentials and Expose them in Environment using Jenkinsfile pipeline

Introduction In this post, we will see how to fetch multiple credentials and…

How to put Code in your blog/article

How to put Code in your blog/article

For programmers, who want to write about their codes. Its often the first…

Solving Jboss Wildfly Oracle JDBC driver problem, with Dockerfile

Solving Jboss Wildfly Oracle JDBC driver problem, with Dockerfile

Assuming your web application is using oracle, and you are deploying your app on…

Docker Push&#58; How to push your docker image to your organization in hub.docker.com

Docker Push&#58; How to push your docker image to your organization in hub.docker.com

Tag the image, by seeing its image id, from docker images command docker tag 04d…

Latest Posts

REST API Design: Pagination, Versioning, and Best Practices

REST API Design: Pagination, Versioning, and Best Practices

Every time two systems need to talk, someone has to design the contract between…

Efficient Data Modelling: A Practical Guide for Production Systems

Efficient Data Modelling: A Practical Guide for Production Systems

Most engineers learn data modelling backwards. They draw an ER diagram…

Deep Dive on Caching: From Browser to Database

Deep Dive on Caching: From Browser to Database

“There are only two hard things in Computer Science: cache invalidation and…

System Design Patterns for Real-Time Updates at High Traffic

System Design Patterns for Real-Time Updates at High Traffic

The previous articles in this series covered scaling reads and scaling writes…

System Design Patterns for Scaling Writes

System Design Patterns for Scaling Writes

In the companion article on scaling reads, we covered caching, replicas, and…

System Design Patterns for Managing Long-Running Tasks

System Design Patterns for Managing Long-Running Tasks

Introduction Some operations simply can’t finish in the time a user is willing…