> Eric .... I realize that O (1) deletion is hard, and don't see a good way around it without changing the implementation ... I just think that the preserving the current C layout may be forcing an even more complicated solution than neccessary.  I am nervous about pushing this to 3.5 because of the complexity.  I agree that a simpler solution should (also?) wait for 3.6.

Noted (and thanks for the feedback).  I'm still comfortable with
moving ahead for 3.5 with what we have.  The code is documented and
structured in such a way that it should be clear what's going on and
relatively straightforward to adjust.  There's a decent chance we will
find a bug or two in corner cases, but nothing at a scale that would
give me pause for a 3.5 release.  Furthermore, the test suite for
OrderedDict is pretty thorough so strict compatibility with the pure
Python OrderedDict allows us to derive a lot of confidence about the C
implementation.

>
> The question about dictheaher.h boils down to: if someone asks whether something is a dictview (not even using "exact", though that isn't an option), an  odictview will say false ... is that really what you want?

Ah.  I misunderstood your question to imply what should be added.
Instead, you were just indicating what is already there.  I don't
think anything needs to be changed though.  Those checks don't pass
for the pure Python OrderedDict so I would not expect them to do so
for the C implementation.