Repo-cmd provides a simple framework for any scripting logic related to a git repository, that should stay outside of the repository. For example, to provide secrets, or to specify some scripting logic that is still half-baked / in some way not yet ready to be shared with fellow programmers.

Of course, one way to achieve this is to simply store a script inside the related git repo, together with an entry in the .gitignore file. However, in my experience this turned out to insufficiently stimulate consistency between different repos.

Additionally, there turned out to be a few features that a framework could quite easily provide:

• print out available commands to run
• provide a convenient note taking CLI (see "Notes feature" section for more info)
• establish a sane skeleton script for new repos

I've personally been using this framework for 2 years before having opened it up in this public Github repo, and it definitely helped me write things down into scripts sooner/more than I had done before. The barriers to do this are lowered simply because of the CLI that repo-cmd elicits. One warning regarding this however; it becomes too easy/inviting to write script logic outside of git repos, pay attention that logic that is ready to be shared should arrive in the git repo itself.

Simply execute the install bash script. This will place the repo-cmd in /usr/bin/. It will additionally install a bash_completion script inside either $BASH_COMPLETION_USER_DIR or if this is not set in ~/.local/share/bash_completion/.

For bash completions to work, make sure you've installed (bash-completion)[https://github.com/scop/bash-completion].

In one of your git projects, add a command (with repo-cmd edit) that launches a Dockerized postgres database:

testdb)
    docker run \
        -p 5432:5432 \
        -e POSTGRES_PASSWORD=somePassword \
        -e POSTGRES_USER=someUser \
        -e POSTGRES_DB=someDatabase \
        postgres
    ;;

(Inside the case).

And then add another command that connects to this database with psql:

db)
    PGHOST=localhost \
        PGPORT=5432 \
        PGUSER=someUser \
        PGPASSWORD=somePassword \
        PGDATABASE=someDatabase \
        psql
    ;;

With these commands defined, it's now possible to start a temporary database inside a terminal that is somewhere inside your git project with repo-cmd testdb. With repo-cmd db you can then connect to this database in another terminal session.

Anything goes. Maybe you want to write some convenience wrapper around API calls you're performing with something like curl or httpie, eg:

login)
    set -u
    username=$2
    password=$3

    response=`http POST https://your_api.com/login <<EOF
{
    "username": "$username",
    "password": "$password"
}
EOF`
    echo "$response" | jq -r '.token'
    ;;
cart-add)
    set -u
    item_id=$2

    token=`repo-cmd login someUser somePassword`
    http https://your_api.com/cart/add/$item_id Authorization:"Bearer $token"
    ;;

repo-cmd always first changes directory to the git project root directory. So, at entrypoint of all of your custom commands, $PWD will always be that git root directory.

repo-cmd additionally provides a way to make notes in 1 file per extension per git project, by executing repo-cmd notes.<extension>.

For example, I use this a lot to write and store one-off SQL queries with repo-cmd notes.sql, and for some projects I may use repo-cmd notes.txt to store findings.