git is a version control system that does a number of things rather “differently.” By now, I suspect most people are familiar with git so I will refrain from focusing too much time here on explaining what git is or how it works. Git does things differently, and to make a long story short, git has no concept of a file rename. To be honest git doesn’t have a lot of sense of individual files at all, but that’s another story. As a result git looks at the contents of the files, and if a new file is similar enough to an old file that was deleted git recognizes this as a “rename,” or a moved filed. It can process file copies in the same manner.
You’d think this would be a problem, but it turns out that it isn’t. In fact it’s a great thing. You can store a bunch of data on a file system, use conventional tools to manipulate, move, and rename those files in the normal manner, and then do a little magic, and git automatically knows everything that it needs to know about what files were renamed and when.
This post is about that magic.
What you need to do is find out the names of the files that existed in
the last commit but have been removed from the file system. If you run
“git rm” on these files and then add everything that remains, git
will be able to pick up the rename operations implicitly, and the next
commit will reflect the current state of the file system.
I’ve thrown this little snippet in the code section of the site: `git-readd </code/git-readd>`_.
I’ve used this process as part of scripts in the past, and it’s great for managing Maildirs with git, but I found myself needing this operation from time to time as I’m manging a git repository. Enter, “git aliases.”
Aliases let you create git sub commands from your own scripts. Add the
following lines to the ~/.gitconfig file on your system, and create it
if necessary:
[alias]
readd = "~/scripts/git-readd"
Now you can run this command in any git repository on your system with the following command:
git readd
Magic!