tl;dr Adds a git identify command to set the current repo's config to a pre-configured identity.
Easily configure different git identities on the same machine.
Have you ever added a one line change to a repository and only realized after pushing that you've added the commit with the wrong name and email? Yeah - me too.
git-identify tries to solve that problem by declaring identities and
their mappings in one place - a ~/.git_identities file.
Pull down the git-identify repository. Then, add the git-identify file to your
$PATH. I've symlinked it into usr/local in the example below but copying will also
work.
git clone https://github.com/ConnorAtherton/git-identify.git && cd git-identify
ln -fs $(pwd)/bin/git-identify /usr/local/bin/git-identifyTo start, create a ~/.git_identities file that follows the example below.
The syntax tries to follow the .git/config style where possible.
[identity:work]
name = Connor Atherton
email = [email protected]
[identity:personal]
name = Connor Atherton
email = [email protected]
[personal]
~/repos/ideas/*
/Users/Connor/repos/personal/*
[work]
/Users/Connor/repos/work/*
A few key points:
- Whitespace is not significant and is stripped out
- Identity declaration order doesn't matter
- The first match always wins
- The globs are case-insensitive (mainly because I use zsh)
Enter into a repo and run git identify to modify the local git
configuration according to the matching identity in the .git_identities file..
Check your config in the target repo to make sure the identity rules are correct.
cd ~/repo
git identify
git config -lSearch for the last occurence of the user.name and user.email fields.
We use BATS for testing. To test, run
bats tests/*.bats from the root of this repo and it will print out the test
results.
How is this better/worse than setting git config manually
It stops you from forgetting to do that as long as you get into the habit of running git identity.