I will take some time in the next week to look at this issue.  Thank you
for bringing these bugs to my attention.  

I did not finish the memoryview implementation, due to other demands on
my time.  I appreciate the efforts of so many to pick up and understand
the PEP and contribute code.   I've noticed some incorrect statements
being passed around, and I apologize for not providing a perspective on
them.   

I hope there is a clear distinction between the buffer protocol and the
memoryview object.   The buffer protocol was thought through fairly
well, but the memoryview object has not received the same attention.  My
mental model of the memoryview object is basically an (enhanced) NumPy
array without the math support.  

It does not suprise me that there are a few remaining issues in the
memoryview object.  However, I'm confident they can be cleared up.   A
few changes to the code were made by others without changing the PEP.  I
accept responsibility for this because people were cleaning up what I
should have already finished.  

All the help is appreciated.  I will review the patches and fix the
problem.