As discussed internally, we'll have a different issue for non-ivarmap solvers where the problem remain.
However, for solvers without ivarmap, it is created each function call, so there the issue likely remains. I think for those solvers, they should also just have an ivarmap if we're going to do any int2bool powered rewriting. I'm setting this PR as ready, since that issue should probably be a different PR.
Originally posted by @hbierlee in #860 (comment)
As discussed internally, we'll have a different issue for non-ivarmap solvers where the problem remain.
Originally posted by @hbierlee in #860 (comment)