A site mirror I made didn't work properly until I commented out the guts of jump_normalized_const so it didn't jump over a www. prefix, and then did work afterwards (although this is oversimplifying as I ended up needing to use https://github.com/mitchcapper/httrack so the project would build, and then had to patch out a couple of regressions it had versus this version).

If the options to treat http:// and https:// URLs as the same, treat www.thing.com and thing.com URLs as the same, and to remove redundant slashes were separate instead of under one umbrella URL Hacks setting, I could have just enabled and disabled the bits I needed.