First of all, the premise "exports > 0" in your example looks wrong
to me.  The deallocation process for the first view should start precisely when it no longer has any exports.

In fact, the check for "exports > 0" is for the case when
memoryview.release() is called from the Python level.


Secondly, even if it did happen (show the code path leading
to that!), BufferError would be set in memory_clear() and the
garbage collector would throw a FatalError, i.e. step 5+ would
not be reached.


Lastly, I don't find it very diplomatic to use language like "Deallocation scheme for memoryview is complex and unsafe. It
crashes with chained memoryviews..." when you don't seem to
a test case or a clear concept of how the alleged bug should
occur.