Add support for is_nan and other floating-point functions

[lovasoa]

Add support for is_nan, is_finite, and other floating-point testing functions

[ISibboI]

Thank you, this looks cool. Two things:

• Can we put these function under the math:: prefix, since they seem to relate to those functions?
• Can we have the is_float function forward the error if the argument is not a float, as opposed to returning false in that case? I think it would make sense to be type-safer here.

[lovasoa]

What purpose does the math:: prefix serve ? is_nan is not really related to math, is it ? Why doesn't round have this prefix ? There is no support for a module system or for namespaces, the math:: prefix seems to just be part of the function name; I'm not sure what its purpose is. I implemented the change anyway... having the function return false on non-float input makes them easier to use, at least in my usecase. I'll switch the behavior to what you suggest, but I'll have to add a new "typeof" function to get the type of a value in order to be able to implement the same behavior in the evalexpr expression itself.

[lovasoa]

@ISibboI : Is there something more you would like me to change before this can be merged ?

[ISibboI]

Thank you very much!

The idea behind putting the functions into modules without having a module system is that we can implement a module system without too large breaking changes.

Having floor, round and ceil as plain functions is likely a mistake.

The typeof function is a good idea!

[lovasoa]

I added the comment you wanted. Does it look ok to you now ?

[ISibboI]

Cool, thanks!

[lovasoa]

Thanks for merging !