It turns out that the key length restriction put in place here was only necessary due to a bug in Firewood, which is now fixed. I think we can reduce the duplicate code, but it would touch quite a bit of merkledb code too. What if after we merge this PR, I refactor all the testing to work for both databases? I think most of the benefit will come once we want to start testing change proofs as well. I worry the scope of refactoring all these tests will creep into implementation
Originally posted by @alarso16 in #4361
Originally posted by @alarso16 in #4361