Description of problem:
When running 0.4.0 tests (with 605a20f applied) on Fedora rawhide on armv7hl (ARM 32-bit), gridData.tests.test_dx.test_write_dx_ValueError test fails:

Version-Release number of selected component (if applicable):
python-GridDataFormats-0.4.0-3.fc29

How reproducible:
Always

Steps to Reproduce:
1. cd GridDataFormats-0.4.0
2. /usr/bin/python2 setup.py nosetests

Actual results:
running nosetests
running egg_info
writing requirements to GridDataFormats.egg-info/requires.txt
writing GridDataFormats.egg-info/PKG-INFO
writing top-level names to GridDataFormats.egg-info/top_level.txt
writing dependency_links to GridDataFormats.egg-info/dependency_links.txt
reading manifest file 'GridDataFormats.egg-info/SOURCES.txt'
reading manifest template 'MANIFEST.in'
warning: no files found matching 'README'
writing manifest file 'GridDataFormats.egg-info/SOURCES.txt'
.............F............S..
======================================================================
FAIL: gridData.tests.test_dx.test_write_dx_ValueError
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/usr/lib/python2.7/site-packages/nose/case.py", line 197, in runTest
    self.test(*self.arg)
  File "/usr/lib/python2.7/site-packages/nose/tools/nontrivial.py", line 60, in newfunc
    func(*arg, **kw)
  File "/builddir/build/BUILD/GridDataFormats-0.4.0/gridData/tests/test_dx.py", line 89, in test_write_dx_ValueError
    return _test_write_dx(nptype=nptype, dxtype=dxtype)
  File "/builddir/build/BUILD/GridDataFormats-0.4.0/gridData/tests/test_dx.py", line 48, in _test_write_dx
    assert_equal(out_dxtype, dxtype)
  File "/usr/lib/python2.7/site-packages/numpy/testing/nose_tools/utils.py", line 411, in assert_equal
    raise AssertionError(msg)
AssertionError: 
Items are not equal:
 ACTUAL: 'double'
 DESIRED: 'unknown'
----------------------------------------------------------------------
Ran 29 tests in 1.106s
FAILED (SKIP=1, failures=1)

Expected results:
Test passes.

Additional info:
It looks like np.longdouble and np.double are different on x86 32bit:

$ arch
i686
$ rpm -q python2-numpy
python2-numpy-1.14.5-3.fc29.i686
$ python2
Python 2.7.15 (default, Jul 15 2018, 06:03:07) 
[GCC 8.1.1 20180626 (Red Hat 8.1.1-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> np.longdouble
<type 'numpy.float96'>
>>> np.double
<type 'numpy.float64'>
>>> 

but they're the same on ARM 32bit:

$ arch
armv7l
$ rpm -q python2-numpy
python2-numpy-1.14.5-3.fc29.armv7hl
$ python2
Python 2.7.15 (default, Jul 15 2018, 05:48:59) 
[GCC 8.1.1 20180626 (Red Hat 8.1.1-4)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy as np
>>> np.longdouble
<type 'numpy.float64'>
>>> np.double    
<type 'numpy.float64'>
>>> 

I'm not sure if this means it's a bug in numpy or the test is making some unwarranted assumption.