cygwin/ARM:pynac.pyx use double precision special functions instead of long double (REVISITED)

[williamstein]

This is a followup to #8847, which was supposed to fix this, but simply didn't.

sage -t  "devel/sage/sage/functions/other.py"               
**********************************************************************
File "/home/wstein/sage-4.4.3/devel/sage/sage/functions/other.py", line 475:
    sage: gamma1(float(6))
Expected:
    120.0
Got:
    119.99999999999997
**********************************************************************
1 items had failures:
   1 of  29 in __main__.example_12

See #8847 for more details. Fixed by #12449.

CC: @SnarkBoojum

Component: porting: Cygwin

Reviewer: Burcin Erocal

Issue created by migration from https://trac.sagemath.org/ticket/9162

[mwhansen]

comment:2

I don't think that #8847 was supposed to fix this. #8847 was to make it so that we get 120.0 instead of 119.99999999999997 on Linux. Cygwin has always given 119.9999...

[kcrisman]

comment:3

Somewhat miraculously, this passes tests on a recent build of mine on XP. Perhaps due to using mpmath now or something?

[kcrisman]

comment:5

But this, once again, does not pass when done by hand. What is going on?

[dimpase]

comment:6

exactly the same problem also pops up on ARM running Ubuntu 11.10.

[burcin]

comment:8

#12449 contains a patch to fix. We should close this as a duplicate.

[jdemeyer]

Reviewer: Burcin Erocal