How I Use: Git

I use Git to track versions of files. Here's how:

Summary

• I create repositories with git init
• I commit with git commit
• I mostly follow semantic commit conventions when writing commit messages
• I push with git push
• I add remote origins with git remote add origin
• If I'm pushing after adding the origin remote, I git push -u origin master

Git Configuration

I have a file ~/.gitconfig. I think it has been mostly auto-populated. Here are the settings I like it to contain:

[core]
	editor = nano
[merge]
	tool = kdiff3

• I prefer nano for editing commit messages
• I prefer kdiff3 for resolving merge conflicts

Commit Messages

I usually write commits like this:

prefix: headline

More details here

http://link.example/some/task/tracker

Some prefixes I use are:

• feat: something new
• fix: something fixed
• test: something test-related
• style: something formatted
• chore: something related to CI or build pipeline

I don't use the refactor prefix.

Rebasing

If I fall behind an upstream remote, I git rebase master. This keeps my Git history clean.

If I try to push and see that I'm behind, I do the above.

If I'm developing a feature, I sometimes make commits like feat: WIP. I sometimes make multiple of these, causing my commit history to look like:

feat: WIP
feat: WIP
feat: WIP 2
feat: WIP

Before publishing my work or asking for review, I squash these commits into one and give them a nicer commit message. I use a command like git rebase -i HEAD~4 (where 4 is the amount of commits we want to change) and change the commits to s for squash. Then, I run git commit --amend to change the message.

TUI

I almost always use tig when I use Git. It's much faster for me than writing the commands. See How I Use: Tig for Git

Secrets

I use git-crypt for storing secrets in Git. See How I Use: Git Crypt for Git