'git commit -a' and 'git add'

I’ve heard or read too many git blog posts/pod casts state that if you create a new file in your local Git repo and you want to shorten the steps on getting it added to the local repository, all you have to do is:

git commit -am "my commit message"

This is not true. If the file has never been added to the repo prior, then you still have to:

git add <file_name>

Why?

The “git commit -a” command is a shortcut to a two-step process. After you modify a file that is already known by the repo, you still have to tell the repo, “Hey! I want to add this to the staged files and eventually commit it to you.” That is done by issuing the “git add” command. “git commit -a” is staging the file and committing it in one step.

If you create a new file, edit it, and issue the “git commit -a” command, you will see something like:

~ > mkdir test
~ > cd test
~/test > git initInitialized empty Git repository in /Users/jasonmeridth/test/.git/
~/test(master) > touch test
~/test(master) > git commit -am "initial commit"
# On branch master
#
# Initial commit
#
# Untracked files:
#   (use "git add <file>..." to include in what will be committed)
#
#       test
nothing added to commit but untracked files present (use "git add" to track)

The “nothing added to commit but untracked files present” is the key comment. Git even suggests using “git add” to track the file. See, it’s user friendly. Just be aware.