Cygwin Binary on Wiki has been updated

[Jerry DeLisle]

Hi Folks,

I have updated the wiki and the binary available there for the Cygwin build.

Regression testing shows one unexpected failure of atan2_1.f90

Given this:

do i = 1, 10
   if(atan(1.0,  i/10.0)  -atan2(1.0,  i/10.)    /= 0.0) &
   &  print *, (atan(1.0,  i/10.0)  -atan2(1.0,  i/10.))
   if(atan(1.0d0,i/10.0d0)-atan2(1.0d0,i/10.0d0) /= 0.0d0) &
   &  print *, (atan(1.0d0,i/10.0d0)-atan2(1.0d0,i/10.0d0))
end do

The results are small but not exactly zero as expected.

$ ./a.exe
  -4.50335449E-08
   1.08094956596804792E-016
   4.03570652E-08
   3.26344853918136835E-017
   2.59082231E-08
   3.02492406123455737E-017
   2.39143301E-08
  -7.53520509877425582E-017
   4.87428657E-08
  -9.36750677027475831E-017
  -2.40507294E-08
   2.21177243187042905E-017
  -1.97886663E-09
  -1.94072188874905294E-017
   1.57936935E-08
  -2.93818788743571702E-017
  -2.46570160E-08
  -4.27175655959288747E-017
   2.18556977E-08
  -3.06287113727155003E-017

Maybe we change the test or XFAIL it?

Jerry