[with patch, positive review] fix dumb error message when modding out by 0: Mod(10,0)

[William Stein]

When doing Mod(n,0), either there should be a useful error message, or one should get n back. The following is no good at all -- one shouldn't get an AttributeError?, which is surely due to a bug.

sage: a = Mod(10,0)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)

    115     cdef IntegerMod_abstract x
--> 116     x = IntegerMod(integer_mod_ring.IntegerModRing(m), n)
    117     if parent is None:
    118         return x

/Users/wstein/build/sage-3.4.2.rc0/local/lib/python2.5/site-packages/sage/rings/integer_mod.so in sage.rings.integer_mod.IntegerMod (sage/rings/integer_mod.c:2730)()
    132     cdef NativeIntStruct modulus
    133     cdef Py_ssize_t res
--> 134     modulus = parent._pyx_order
    135     if modulus.table is not None:
    136         if PY_TYPE_CHECK(value, sage.rings.integer.Integer) or PY_TYPE_CHECK(value, int) or PY_TYPE_CHECK(value, long):

AttributeError: 'sage.rings.integer_ring.IntegerRing_class' object has no attribute '_pyx_order'

[Minh Van Nguyen]

based on Sage 4.1.alpha1

[Karl-Dieter Crisman]

Applies fine to 4.1.alpha1 and does what it says. I like the second doctest. Unless something goes wrong with docbuild due to the #, all is well.