Starting from this commit https://github.com/clojure-emacs/clj-refactor.el/commit/363b95c5d2855abc93ac011e9adc778cf7a773e5 we're giving preference to unit tests - more integration-y tests were painful to maintain. So as long as you can extract various logic to pure defuns, and have those defuns tested, your PR will be fine :)

I'll review it tomorrow or so

Ok, cool, I can add some of those. It definitely changed existing behavior though, so I guess pre the code review the main question is does the change make sense? Worried it helps a lot for my use cases, but guessing by the couple failing integration tests there are a couple use cases I may have missed. Anyway, it's more of a feature spike at the moment, happy to tune all the other requirements if it's a suitable approach.

I took a very quick look so I still haven't grasped what it's about in my local copy I use ido-completing-read so I don't like the default behavior either :) but, I try to never inflict breaking changes, be it from UX or API point of view. So changes would have to go through defcustoms

Yea I mean if we want the defcustom route, I can just make an alternate version of cljr-slash for folks to bind.