Git Stash Changesets

Pretty frequently I start writing some code, when I realise there’s another change I need to make before I can continue. I like to make lots of small atomic changes to a code base because it lets me test more quickly and catch errors earlier. I used to do this by saving my changes in a temporary file, but this was clunky. A better way is with git stash.

But git stash reverts all files; and very often I want to keep some, especially configuration parameters. However there’s a way to stash just the files added for commit; use git stash --index. The name comes from the index being the list of changes that are to be added for the next commit. I use this in Emacs via magit which lets me easily stash chunks.

The only downside of this is it won’t let me pop the stash until the file is at the latest commit. This means if I want to combine the changes in a single commit I have to commit, then pop the stash and amend the commit. Not awful, just a little finicky. But it’s a pretty safe way to work.

Another alternative to stashes is branches. It’s very easy to make a temporary branch to store some work and move around. But stashes require a bit less context switching, and are convenient to use.