Fix various bugs in DiscordBitSet

[lukellmann]

• WIDTH was defined to the width of a Byte not a Long, resulting in bugs in size, get and set

• binary just did some completely wrong things that I can't really describe

• hashCode could return different values for equal DiscordBitSets (when they had different amounts of trailing zeros in the data array), breaking its general contract

• equals was comparing this with this and not this with other, which resulted in any two DiscordBitSets considered equal

• contains would erroneously return false if other did have more trailing zeros but would otherwise be contained

• set wasn't growing data correctly and also didn't unset bits properly

[lukellmann]

please tell me if you happen to find any other bugs or if you think my fixes still include bugs

[DRSchlaubi]

Don't know enough about this to review this properly

[HopeBaron]

The re-write is perfectly fine much easier to read as well.
But the initial reason we used a byte size is to save ourselves from allocating non-necessary bits

[lukellmann]

But the initial reason we used a byte size is to save ourselves from allocating non-necessary bits

yeah but since the implementation uses a LongArray rn, setting WIDTH to Byte.SIZE_BITS was simply wrong

[HopeBaron]

You are right, let's just keep it long then

[lukellmann]

did some last small changes, see here

[HopeBaron]

They look good, if there's nothing else you want to do on this PR we can merge

[lukellmann]

i'm done here 👍