Open mccalluc opened 3 weeks ago
Error is coming from R build... probably not from this PR?
Can file an issue if we see this again. Will update and re-run CI.
(leaving open until @Shoeboxam has chance to review)
The reason why it's taken me so long to review this PR is because I find the new code to be very complicated, and I feel that it could be simpler, but I don't know what to recommend off-hand. I've tinkered with it some myself, but haven't had uninterrupted time to take a good look.
Not super confident about this approach, but
check_similar_scalar
really seems to be the center of things: This gets called from many different points, so doing it outside this function seemed like a piecemeal solution.This does fix
test_int_data_laplace_param
and I've added a lower level conversion test, but there is a change in behavior in binary search, but maybe that's ok?Draft for now, but feedback welcome.