My Git Second Brain

Published 12-15-2023

These are my personal notes on git. Each section starts with a problem or goal, and is further broken down into specific scenarios. This way, the notes are structured “Just in Time” (like asking a friend for help). Instead of “Just in Case” (like reading the manual start-to-finish before driving the car).

Table of Contents

• Rapid fire tips
• “How do I safely rebase (never fear rebase again)”
• “I messed up. How do I fix, revert or undo this?”
• “How do I use git effectively for me (and my team)”
• Worktrees

Rapid Fire Tips

• git pull translates to git fetch + git merge (or git rebase given -r)
• git pull -r origin main to rebase the latest change from main (with 1 command)
• git fetch is still useful on its own. Before checking out a teammates remote branch, run git fetch
• git fetch → git reset --hard origin/main to “force pull” (use with caution)
  • Whoops. git reset HEAD@{1} (you can undo a bad reset with… reset)
• git log -g instead of git reflog. Personal preference.
• git checkout -b temporary-branch before commit / push if you’re worried about breaking something
• git reset --soft put your commits back to the staged state. Now you can commit multiple previous commits as one well-written commit. See examples
• git push will automatically point to git push ORIGIN FEATURE-BRANCH given this config git config --global push.autoSetupRemote true
• git push --force-with-lease force pushes but prevents you from overwriting others work
• git clone repo_name --bare + git worktree add your_branch is great for devs who frequently switch branches
• You should try Conventional commits. They make your commits easy to read and can automate tags/releases/versioning
  • Oh wow. What coincidence. Totally unexpected. I wrote a CLI that helps you write conventional commits - Better Commits Github

How do I safely rebase (never fear rebase again)

A few assumptions

• You’re working on a feature branch
• origin is the name of your remote
• main is the name of your default branch

Is it safe to rebase (strict) - Is your branch only local? (not public)

• Yes - Safe to rebase ✅
• No - Not safe to rebase ❌

Is it safe to rebase (simple version) - Are you the only person working on your branch?

• Yes - Safe to rebase ✅
• No - Not safe to rebase ❌

If you work on a team where developers checkout their own feature branch, and only they work on it. Go for the simple version. If developers commonly work on each other branches or share long lived branches. Go for the strict version.

How can you guarantee it’s safe to rebase 100% of the time? git checkout -b new-branch-name. Because now your branch only exists locally… until you push of course.

Rebase Checklist

1. git branch → git log verify feature branch and existing commits:
2. git add → git commit → git push origin feature-branch if you have changes, commit
3. git pull -r origin main pull
4. If you have conflicts, fix → git add → git rebase --continue
5. git log verify your work
6. git push --force update your feature branch

Rebase Checklist Explanation & Tips

1

Make sure you’re on a feature/developer branch. Doing this directly on a main/release branch is a big no-no.

If you like to be thorough (I’m a fan) git log so you know what the log looked like before your rebase.

Concerned about making a mistake? git checkout -b temporary-branch. Now any mistakes will be isolated to this private temporary-branch. Skip this step if you’re feeling confident (Don’t worry, I believe in you).

2

If you have changes then stage, commit and push. git add, git commit, git push. Very important. If you don’t make your own commit before rebase you may accidentally merge the work of another developer into your work and restarting from a mistake becomes more difficult.

Multiple Commits & Rebasing (Situational)

Given multiple commits (and possible merge conflicts). Squashing your commits first in this scenario saves you a lot of time because merge conflicts are compared against each commit.

3

git pull -r origin main. Pull -r means git fetch + git rebase, this guarantees you pull the latest commit.

4

If you have merge conflicts you’ll likely use your editor to resolve them. After resolving, remember these commands. git add will stage your resolutions, git rebase --continue will move to the next step. If you make a mistake, you can restart via git rebase --abort

5

Double check your work. At this point your branch should resemble main + your commits replayed on top.

6

git push --force. Don’t fear the force flag. In fact, you need to use force here.

Tip #1 Alternatively, you can use force-with-lease instead of force as a safer version of force. This prevents you from accidentally overwriting someone elses work. If you followed the steps, that shouldn’t be an issue, but it doesn’t hurt to use.

Tip #2 git config --global push.autoSetupRemote true this allows you to type git push instead of git push origin name-of-your-branch-you-have-checked-out

An Example

git log # Double check existing commits

# Commit changes (if any)
git status 
git add .
git commit -m "My sweet new feature"
git push origin my-feature-branch
# End changes

git pull -r origin main

# Fix Conflicts (if any)
git add .
git rebase --continue
# End Conflicts

git log # Verify we did everything correctly
git push origin my-feature-branch --force

I messed up. How do I fix, revert or undo this?

I want to undo my commit, but I don’t want to lose those changes

• git reset --soft HEAD~1 resets 1 commit back to staged

I want to remove a file from the staged state

• git restore --staged PATH_TO_FILE

I committed a file to my remote feature branch, now I want to remove it

git reset --soft HEAD~1
git restore --staged PATH_TO_FILE_TO_REMOVE
git commit
git push --force-with-lease # Your feature branch

I want to undo my commit, and I want to completely remove those changes

• git reset --hard HEAD~1 resets 1 commit completely
• To remove the commit from the remote, follow with git push --force. Do this with caution and follow the same rebase rules above when doing so.

Help! I used reset --hard and I meant to use reset --soft

• git reset HEAD@{1} to undo your latest git operation

How does this work? Run git log -g or git reflog. Both commands show a reference log which include entries such as reset. For example, if the very last command you ran was git reset --hard origin/main you may see: Reflog: HEAD@{1} Reflog message: reset: moving to origin/main.

Reflog is your go to any time you make mistake but don’t have a specific commit to fall back to.

I want to remove all changes and “force pull” from the remote

• git fetch → git reset --hard origin/your-branch-name

I have multiple commits on my remote feature branch, I want to combine them into 1 commit

• First, follow the rebase rules. i.e. make sure you can safely force push without losing someone elses work.
• git log count the number of commits you want to squash (using 4 for the example)
• git reset --soft HEAD~4
• git commit -m "your new message"
• git push --force

Git Clean Guide (Removing unwanted files)

I want to remove untracked files

• git clean -f

I want to remove untracked folders/directories

• git clean -fd

I added something to my .gitignore that I previously commited, now I want to remove it.

• git rm —cached `git ls-files -i -c —exclude-from=.gitignore`

I want to remove changes in my working directory

• git restore FILE_NAME

I want to remove ignored files

• git clean -fX

Notice the capital X, you probably want capital. Lowercase x will remove files like .environment.

I was working off of another developers branch, they merged to main. Now I have conflicts with their merge commit when I try to merge to main.

There is very likely a more efficient method to do this, but as of today, here’s how I handle it.

git log # Copy each commit hash you've made
# abcd1235 
# efgh6789
# lmno1234 → Commit you started from

git checkout main
git pull -r origin main # Now you should have their changes
git checkout -b your-new-feature-branch

# Apply from oldest to most recent
git cherry-pick efgh6789
git cherry-pick abcd1235

# Push
git push origin your-new-feature-branch

💡 You can simplify this process by having 1 commit. If you have multiple commits, see previous section “I want to combine them into 1 commit”

How do I use git effectively for me (and my team)

Branch naming conventions

Any variation of this can help define intention. It’s also useful for automating actions, tags, versions and releases

• username/type/TICKET-description
• username/type/description
• type/description

Commit message conventions

You can find additional examples and explanations here: Conventional Commits Summary

feat(tools): #3 infer ticket/issue from branch

If commit_type.infer_value_from_branch is enabled,
will attempt to infer ticket/issue from branch name.
Updated documentation with details.

Closes: #3

My workflow - New branch checklist

1. checkout default git checkout main
2. update main git pull -r origin main
3. install dependencies (example) npm ci

Better Commits is a CLI to achieve all of the above. It’s well documented, flexible, easy to install and I use it every day. It can help you (and your team) create a consistent branch / commit workflow and minimize mistakes across users without enforcing PR lint rules or pre-commit hooks.

Using Worktrees

Manage multiple working trees attached to the same repository source

In other words, they allow you to have multiple commits checked out at the same time in the same repository.

git clone --bare YOUR_REPO it’s recommended to use a bare repository with worktrees.

How do I Create Worktrees

git worktree add main This will checkout a new worktree (named main) at main’s HEAD.

This is a little confusing, because the worktree name and the branch name will match. — If the branch exists, it will checkout that branch. If it doesn’t exist, it will create a new branch that starts at the HEAD of main/master.

Examples

# main branch
$ git worktree add main
Preparing worktree (checking out 'main')
HEAD is now at c69d66d fix: improve docs (#9)

# existing feature branch
$ git worktree add about-me
Preparing worktree (checking out 'about-me')
HEAD is now at 7d6a74b Fix paths in prod

# new feature branch -- notice commit is same as main
$ git worktree add does-not-exist
Preparing worktree (new branch 'does-not-exist')
HEAD is now at c69d66d fix: improve docs (#9)

Running ls you can see each worktree is represented by a folder, which are essentailly seperate checkouts with their own state. Concrete example, if you run npm install in each folder. Each worktree will have it’s own node_modules.

How Do I Delete Worktrees

git worktree remove NAME

“fatal: contains modified or untracked files, use —force to delete it”

git worktree remove --force NAME

How Do I Use / In My Worktree Name

If you’re using a branch naming convention, e.g. something like this username/type/ticket-description. Each / will be transformed into a nested folder (which may not be desirable).

Using better-commits, you can run the better-branch command, which supports Worktrees.

• It’s a handy way to create worktrees while still using branch naming conventions

Alternatively, git worktree add WORKTREE_NAME -b BRANCH_NAME to specify a different worktree / branch name manually.