[KScl]

What is this fixing or adding?

A combination of two bugs that seem to have been introduced when converting the upstream code to Python:

• Distribution weights were being doubly subtracted from the given random number; this resulted in Blue Crystals being picked for SM bosses far more often than they should, as well as making it impossible to pick Red Crystals and Pendants for SM bosses. (n - (12 + 15 + 15) where n is between 0 and 35 is always less than 0)
• In the Distribution constructor, passing in another Distribution attempted to turn the newly instantiated Distribution into a copy of it... by overwriting self, which doesn't work. After leaving __init__, that copy (and the changes made to it afterwards by other parameters) would be tossed out, and the result would always be the default Distribution weights. This was masked by the above bug; fixing that one without fixing this one resulted in occasionally attempting to place three Red Crystals, two Green Pendants, etc.

How was this tested?

Test generations showing Red Crystals and rarely, Pendants being placed on SM locations, where they were not before

[KScl]

Relevant upstream code: WorldState.cs:75-90

@lordlou

[ScipioWright]

The math makes much more sense now (I'd prefer if it was just getting a random number then figuring out if it's within a range directly instead of subtracting each time but it's not worth changing it that much), and the other bug here was understandable.

[lordlou]

Oh this was indeed all wrong. Thanks for helping, much appreciated!