Detze

I see now. There's definitely a problem to resolve here. There's already code that hard-assumes that containment in a hit window is always judged using `

I've performed fairly extensive testing aiming to verify the matching of the proposed hit windows against the ground-truth. With this PR, all legacy "true" hit window values are half-integers, so...

The remaining mania replay hit result count mismatches are due to two disparities not related to legacy hit window values: Disparity 1  Consider the third column from the right....

Distinctness of beatmap backgrounds is useful. Instead of replacing the backgrounds with non-unique default image, blank image, or seasonal backgrounds, they could be replaced with some deterministic unique replacement image...