RFC: Add math.isnan, math.isinf and math.isfinite to Math Library

[FeudalDiana]

Rendered

Proposal to add:

math.isnan(x: number): boolean
math.isinf(x: number): boolean
math.isfinite(x: number): boolean

[jackdotink]

I don't think both isinf and isfinite are needed.

[aatxe]

Can you please linewrap the markdown file so that we can leave comments in a targeted way? 80 or 100 or even 120 columns is fine.

[FeudalDiana]

@aatxe Done 😄

[vegorov-rbx]

It's very common for language math APIs to have both isinf and isfinite.

[deviaze]

Could we also support vectors here?

• math.isnan(v: vector) returns true iff any component of the vector is NaN, similar to how v ~= v already works.

• math.isinf(v: vector) returns true iff any component of the vector is +/- math.huge

• math.isfinite(v: vector) returns true iff every component of the vector is finite

[FeudalDiana]

Could we also support vectors here?

• math.isnan(v: vector) returns true iff any component of the vector is NaN, similar to how v ~= v already works.
• math.isinf(v: vector) returns true iff any component of the vector is +/- math.huge
• math.isfinite(v: vector) returns true iff every component of the vector is finite

Feels like this would be better to implement inside of the vector library, going off existing convention (math.max, vector.max etc). Perhaps in a separate RFC?

[aatxe]

Yeah, I don't think we need to talk about vector support for functions in what is effectively the number library