[with patch, with positive review] Solaris 10/Sparc: Fix numerical noise issues in doctests

[mabshoff]

There are a couple doctests on Solaris 10/Sparc that fail due to numerical noise. Fix it.

A patch is coming up.

Cheers,

Michael

[mabshoff]

This patch fixes the following two doctest failures on Solaris 10/Sparc:

sage -t  "devel/sage/sage/modules/free_module_element.pyx"  
**********************************************************************
File "/home/mabshoff/build-3.4.1.rc1/sage-3.4.1.rc1-mark-gcc-4.3.3/devel/sage/sage/modules/free_module_element.pyx", line 505:
    sage: vector(RDF, {1:pi, 1000:e})._sage_input_(SageInputBuilder(), False)
Expected:
    {call: {atomic:vector}({atomic:RDF}, {dict: {{atomic:1}:{atomic:3.1415926535897931}, {atomic:1000}:{atomic:2.7182818284590451}}})}
Got:
    {call: {atomic:vector}({atomic:RDF}, {dict: {{atomic:1}:{atomic:3.1415926535897931}, {atomic:1000}:{atomic:2.7182818284590455}}})}
**********************************************************************
1 items had failures:
   1 of  10 in __main__.example_8

and

sage -t  "devel/sage/sage/rings/real_double.pyx"            
**********************************************************************
File "/home/mabshoff/build-3.4.1.rc1/sage-3.4.1.rc1-mark-gcc-4.3.3/devel/sage/sage/rings/real_double.pyx", line 727:
    sage: sage_input(RDF(-e), verify=True, preparse=False)
Expected:
    # Verified
    -RDF(2.7182818284590451)
Got:
    # Verified
    -RDF(2.7182818284590455)
**********************************************************************

Cheers,

Michael

[craigcitro]

Yep, that's some numerical noise.

[mabshoff]

Merged in Sage 3.4.1.rc3.

Cheers,

Michael