Five Git Habits I Stole From Better Engineers

I’ve read the same “advanced Git” blog posts you have. Rebasing poetry, reflog rescue missions, submodules (never again). In real jobs, tiny habits beat trick shots — the ones I could do on a bad Tuesday are the ones still here.

1. Commit messages with a verb in the present tense

Not because a linter yells — because future me reads git log like a diary.

Bad: wip
Better: fix: handle empty cart in checkout API

2. `git add -p` for everything non-trivial

I’m allergic to shipping debug print()s. Staging hunks forces me to look at the diff. Yes, it’s slower. So is a revert at midnight.

3. Branch names that include the ticket

feat/JIRA-4429-oauth-scopes looks ugly. When someone asks “where’s that fix?” I can grep history in five seconds.

4. Pull with intention

git pull --rebase on feature branches keeps history readable. I don’t fight about merge vs rebase on main — team wins — but on my branch, I like a straight story.

5. One logical commit before opening a PR

Squash locally or on merge, I don’t care — but I stop dumping “address review” as seven micro-commits unless the team wants that granularity.

Git branch visualization concept — Branches are temporary; the story you leave in history is not.

The habit I’m still bad at

Deleting merged branches. My local repo is a graveyard. If you have a one-liner you swear by, I’m listening.

Aliases that survived my dotfile purge

git config --global alias.lg "log --oneline --graph --decorate -20"
git config --global alias.unstage "reset HEAD --"

git lg is the only “pretty log” I need day to day. Everything fancier I forget under stress.

When the team standard disagrees with my prefs

On main, I follow house rules — merge commits, squash merges, whatever ships. On my feature branch, I rebase to keep review readable. Fighting about git theology in PR comments is never the highest-value argument in the room.

`git bisect` paid for itself once

Binary search between good and bad commit found a regression in six checkouts instead of me guessing across two weeks of history. I don’t bisect weekly, but when production is weird and tests are green, it’s the adult in the room.

None of this needs a conference talk. It’s the boring stuff that keeps a team from hating the repo — which might be the whole point.