[Rd] Accuracy of qt for df=1 (PR#9804)

[mwelinder at gmail.com]

Full_Name: Morten Welinder
Version: 2.2.0
OS: Linux
Submission from: (NULL) (216.223.241.229)


The qt function for df=1 is implemented as...

    q = - tan((P+1) * M_PI_2);

Adding 1 kills accuracy for P near 0.  A better way is to use

   q = cot(P * M_PI_2);
or
   q = 1/tan(P * M_PI_2);