> atan(0,0) is not well-defined (more precisely, there is no unique choice
> within the interval [0,360°)) - if the point (0,0) is written in polar
> coordinates (0*cos theta, 0*sin theta), then any angle theta can be used.
>
>
> If you intend to support (arctan2 0 0), it may be good to mention
On 17/02/2025 8:03, Adam Faiz wrote:
+ (cond ((zero? x)
+ (let ((sign (if (positive? y) + -)))
+ (sign 90)))
atan(0,0) is not well-defined (more precisely, there is no unique choice
within the interval [0,360°)) - if the point (0,0) is written in polar
coordinates (0*cos t
This revised patch documents the limitations of this module and adds more unit
tests.
I've also changed the functions to short-circuit only results that are
rational, since
the exactness property is already lost in any case when the square roots are
irrational.
>From 5eb3be7ce0781068d8e0030c4e5
