SLARTG generate a plane rotation so that

    [  CS  SN  ]  .  [ F ]  =  [ R ]   where CS**2 + SN**2 = 1.
    [ -SN  CS  ]     [ G ]     [ 0 ]

 This is a slower, more accurate version of the BLAS1 routine SROTG,
 with the following other differences:
    F and G are unchanged on return.
    If G=0, then CS=1 and SN=0.
    If F=0 and (G .ne. 0), then CS=0 and SN=1 without doing any
       floating point operations (saves work in SBDSQR when
       there are zeros on the diagonal).

 If F exceeds G in magnitude, CS will be positive.

          F is REAL
          The first component of vector to be rotated.

          G is REAL
          The second component of vector to be rotated.

          CS is REAL
          The cosine of the rotation.

          SN is REAL
          The sine of the rotation.

          R is REAL
          The nonzero component of the rotated vector.

  This version has a few statements commented out for thread safety
  (machine parameters are computed on each entry). 10 feb 03, SJH.