Fix mania GenerateHitResults() bugs

[Syriiin]

There are a couple bugs:

1. An odd number of Perfect/Great hits would miss a hit and throw off following calculations.
2. Oks were duplicated as Mehs

This function is quite hard to parse so it's not surprising these slipped in.
I think these variables could do with some renaming, but that's beyond the scope of this fix.

On a side note, my gut says it should just assume all Greats are Perfects rather than splitting them down the middle.
Usually one would expect that omitting a parameter would set it as the best possible value.

[bdach]

On a side note, my gut says it should just assume all Greats are Perfects rather than splitting them down the middle. Usually one would expect that omitting a parameter would set it as the best possible value.

The reason I did this originally is that hitting all perfects seemed - excluding the easiest maps - rather unrealistic. Usually when someone isn't providing exact difficulty numbers they're providing raw accuracy so extrapolations need to be made.

And yeah that wouldn't matter one bit in score V1 but it does actually matter a bit now that perfects are scored 305 rather than 300 when it comes to accuracy at least.

[Syriiin]

The reason I did this originally is that hitting all perfects seemed - excluding the easiest maps - rather unrealistic. Usually when someone isn't providing exact difficulty numbers they're providing raw accuracy so extrapolations need to be made.

Yea in practice for real plays it makes sense, it's just the expectation I think most people would have from difficulty calculators though.
To get the max pp from a score, you just give the beatmap with no parameters. 🤷‍♂️
Needing to explicitly pass a --greats 0 (especially when it's divergent from existing patterns) just feels wrong.
If someone explicitly passed accuracy as --accuracy 99 or something, I can see splitting the difference between greats and perfects as more justifiable there.
Anyway, off topic!