On Tue, Nov 26, 2019 at 02:49:12PM -0700, Stefan Sperling wrote:
> This is a rather large diff. I don't see a good way to split it up
> since tree entries are used all over the place. The regress tests
> are passing and manual testing of tog suggests that nothing breaks,
> though I may have missed something.

This is now in the repo. Please let me know if anything breaks.