That's a really interesting question. I tried to think about that when I was writing this blog post and couldn't really think about anything. Perhaps setting some permissions on a file or running `chattr` / `setfacl` is the best idea I had, but I omitted any suggestions because I've never actually used a post-checkout hook in anger.

Checking out submodules.

Put

    git submodule sync && git submodule update --init --recursive

into `.git/hooks/post-checkout` and never again wonder why your code doesn't work after switching branches (because you forgot to update submodules).

Also put it into `post-rewrite` so that the same works for `git rebase`.

In one repo I use the post-checkout script is "npm install" so I always get the package versions for the branch I just checked out. One less thing to worry about.

I'm stretching here, but maybe compiling assets with webpack and then HUPing your web server? So when you switch between branches, your local server is always ready to serve up the version you've got on disk?

Notifications. Not for people, but for machines: "checkout is complete on X, and you're subscribed to this event, now do something based on that." If you publish on any event, it's the client's responsibility to choose which it wishes to subscribe to.

I use a set of pre/post hooks to save uid & permission data for a repo that contains www & system data.

Pulling a dependency project when a dependent project is checked out maybe