numpy.dtype.char can return different values for different platforms
NITFReader has a reference to a NITFDetails which contains a reference to the file, which may have been opened in the __init__. This open file causes problems in the unit tests because they delete the file before the file object is deleted.
A couple tests failed on Windows.
numpy.dtype.charcan return different values for different platformsNITFReaderhas a reference to aNITFDetailswhich contains a reference to the file, which may have been opened in the__init__. This open file causes problems in the unit tests because they delete the file before the file object is deleted.