p_atan2.c is b/a or a/b?

[twocs]

Looking at the relevant lines of p_atan2.c at https://github.com/parallella/pal/edit/master/src/math/p_atan2.c

• Calculates the inverse tangent (arc tangent) of b/a. Stability not guaranteed
• for 'a' values near 0. Results are in the range of -pi to pi.
• @param a Pointer to denominator input vector
• @param b Pointer to numerator input vector
  void p_atan2_f32(const float *a, const float *b, float *c, int n)

Contrast this with atan2 ISO/IEC 9899:1999 (E)
The atan2 functions compute the value of the arc tangent of y/x, using the signs of both
arguments to determine the quadrant of the return value. A domain error may occur if
both arguments are zero.
double atan2(double y, double x);

In my impression, the signature of atan2(y, x) becomes atan(y/x) with some edge conditions.
However, the signature of p_atan2(a, b, ...) becomes atan(b/a) [or equivalently p_atan2(y, x, ...) becomes atan(x/y)].

Isn't this inverted? Is it intentional?

[jar]

@twocs
It's obviously inverted and it probably wasn't intentional. Breaking an API should generally be avoided, but this software is still under development, so you should put in a pull request to patch it.