Workaround for Math variadic functions

[aspeddro]

This ensures that the return for an empty array does not return -Infinity

TODO:

• Update CHANGELOG

See #53

[glennsl]

Might it be better to just remove the @variadic? (edit: sorry, didn't mean to include this)

[glennsl]

I'm not sure the float versions ought to be overridden. Math.min([]) does indeed return 0, and Math.min() returns Infinity, but Math.minMany([1, 2, 3]) maps to Math.min(1, 2, 3), not Math.min([1, 2, 3]), and so it seems more consistent to have Math.minMany([]) map to Math.min(), not Math.min([]).

These are also edge cases that are very unlikely to occur in practice, since variadic calls require syntactic arrays there's little chance of it happening by accident. I'd suggest the aim of this is restricted to just making the types sound.

[aspeddro]

Might it be better to just remove the @variadic?

If you remove @variadic the compiler passes an array. Math.min([1, 2]) returns NaN

Playground

[glennsl]

If you remove @variadic the compiler passes an array. Math.min([1, 2]) returns NaN

Ah, sorry, I didn't mean to include that. I checked after writing it, but apparently forgot to remove the comment. Please see the other comments instead :)

[aspeddro]

I was wondering and not sure if we should return 0.0. Maybe let returns Infinity.