When starting a new feature, I normally checkout a branch from
master. Before I do this, I always run:
$ git fetch $ git reset --hard origin/master
It’s a habit and I realised that I’m not sure why I do it. Is
git fetch required? How does this differ from
git pull? What’s the difference between
git pull and
git reset --hard origin/master? I decided to find out, in this blog post.
The answer is summarised in this StackOverflow answer. Read on for a more detailed explanation.
git fetch downloads commits, files and branches from the git remote. You’ll need to use this command to get the latest changes that others have made. You’ll also need to use it to checkout a new branch that someone else has pushed.
git pull does two things:
git fetch and then
git merge origin/<branch>. This is useful if someone else has made new commits, on your branch, and you want to merge them into your branch. Git will attempt to auto-merge any local changes. If they cannot be resolves, it will result in merge conflicts.
Sometimes a branch has diverged from origin so much, that it doesn’t make sense to try to resolve all the conflicts. In this case, it’s better to just reset your local branch to whatever is on origin. To do this, you need to fetch first and then run
git reset --hard origin/<branch>.
I can quit my annoying habit of running
git fetch and
git reset --hard origin/master every time I checkout a new branch. As long as
master is clean, it’s fine to run
Subscribe to get articles like this in your inbox, every week.
You'll get my latest blog posts as well as article, book and podcast recommendations. All about tech.