Allow arbitrary strides and zero- and one-sized dimensions

tests/memoryview/relaxed_strides.pyx

+# tag: numpy
+# mode: run
+
+"""
+Test accepting NumPy arrays with arbitrary strides for zero- or one-sized
+dimensions.
+
+Thanks to Nathaniel Smith and Sebastian Berg.
+
+See also:
+
+    Mailing list threads:
+      http://thread.gmane.org/gmane.comp.python.cython.devel/14762
+      http://thread.gmane.org/gmane.comp.python.cython.devel/14634
+
+    Detailed discussion of the difference between numpy/cython's current
+    definition of "contiguity", and the correct definition:
+      http://thread.gmane.org/gmane.comp.python.cython.devel/14634/focus=14640
+
+    The PR implementing NPY_RELAXED_STRIDES_CHECKING:
+      https://github.com/numpy/numpy/pull/3162
+
+    Another test case:
+      https://github.com/numpy/numpy/issues/2956
+"""
+
+import numpy as np
+
+def test_one_sized(array):
+    """
+    >>> a = np.ascontiguousarray(np.arange(10, dtype=np.double)[::100])
+    >>> test_one_sized(a)[0]
+    1.0
+    >>> a = np.arange(10, dtype=np.double)[::100]
+    >>> test_one_sized(a)[0]
+    1.0
+    """
+    cdef double[::1] a = array
+    a[0] += 1.
+    return array
+
+def test_zero_sized(array):
+    """
+    >>> a = np.ascontiguousarray(np.arange(10, dtype=np.double)[100:200:10])
+    >>> a = test_zero_sized(a)
+    >>> a = np.arange(10, dtype=np.double)[100:200:10]
+    >>> a = test_zero_sized(a)
+    """
+    cdef double[::1] a = array
+    return a
\ No newline at end of file