The normalization of `new Decimal(-Infinity)` breaks some operations and may be incorrect

[James103]

The normalization of `new Decimal(-Infinity)` breaks some operations and may be incorrect

[MathCookie17]

I am of the opinion that -Infinity should be {-1, Infinity, Infinity}, but break_eternity was designed with -Infinity as {-1, -Infinity, -Infinity}, so that's what it normalizes to - it just wasn't even registering as -Infinity as all when I had it normalizing to {-1, Infinity, Infinity}. I don't want to cause a breaking change, so despite my preferences, it'll remain as {-1, -Infinity, -Infinity}... so I'll probably need to make a bunch of special cases to fix it in the functions. I'll look into this at some point.

[MathCookie17]

I did an immediate fix for the Eternal Notations issue (i.e. it now detects infinities in a way that sidesteps the break_eternity bug, so -Infinity works there now. Figured I'd note this here since this was kind of an Eternal Notations issue too, though I haven't looked at the actual break_eternity bug yet)

[MathCookie17]

...on second thought, there are just too many strange cases that {-1, -Infinity, -Infinity} causes. I'm changing it to {-1, Infinity, Infinity} like it always should have been.