Several complicated bug fixes (v2.0.0)

[MathCookie17]

• Added a parameter to lambertw allowing you to calculate the non-principal branch
• Changed all cases where a constant or parameter is returned directly, to reduce the likelihood of accidental mutation
• Changed how tetration for bases <= e^1/e works: now the property x^^(n + 1) == x^(x^^n) always holds, and small bases with large payloads behave correctly.
• Fixed f_gamma

[MathCookie17]

This update fixes #159 and #152 (EDIT: and #162). It also fixes #86, but it doesn't fully address it - it fixes the bug that began issue #86, but it does not do the nocopy reoptimization stuff, so it's your call on whether to leave 86 open or not.

There's a lot of changes in this, so perhaps there are more unit tests that should be run that I didn't think of? If there are, feel free to add whatever tests you deem necessary - my changes may have broken something I'm unaware of... and if you need me to explain why I made a particular change, then please ask.

Also, when pushing these commits, I got the message "This diff contains a change in line endings from 'LF' to 'CRLF'.". I think that's just a weird GitHub thing, but I figured I'd let you know in case it's actually a problem.

[Patashu]

Ok I'm evidently not going to get around to this repository any time soon and it's getting a decent amount of activity, so I'll add you as a collaborator. Ping me (email, twitter or discord works) any time you want me to npm publish. Enjoy your newfound powers!

[MathCookie17]

OK. I'm not going to merge this request just yet because I want to fix #168 first, but I'll try to get that done in the next few days.