a much more modest outcome than what we started with.   The
work of create_row_processor() for ColumnProperty objects
is essentially done at query setup time combined with some
lookups in _instance_processor().
- to allow this change for deferred columns, deferred columns
no longer search for themselves in the result.   If they've been
set up as deferred without any explicit directive to undefer them,
then this is what was asked for.  if we don't do this,
then we're stuck with this performance penalty for all deferred
columns which in the vast majority of typical use cases (e.g. loading
large, legacy tables or tables with many/large very seldom
used values) won't be present in the result and won't be accessed at all.

in callcounts specific to the "expiration" of objects, as in
the "auto expire" feature of :meth:`.Session.commit` and
for :meth:`.Session.expire_all`, as well as in the "cleanup" step
which occurs when object states are garbage collected.
fixes #3307

:class:`.Query` before it fetched results, particularly when
row processors can't be formed, the cursor would stay open with
results pending and not actually be closed.  This is typically only
an issue on an interpreter like Pypy where the cursor isn't
immediately GC'ed, and can in some circumstances lead to transactions/
locks being open longer than is desirable.
fixes #3285

N occurrences of a parameterized string.   This allows parameterized
warnings that can refer to their arguments to be delivered a fixed
number of times until allowing Python warning filters to squelch them,
and prevents memory from growing unbounded within Python's
warning registries.
fixes #3178

- an existing state shouldn't need its load_options/load_path updated;
it should maintain those from its original Query source.  there's no
tests that check this behavior

that I'd like to start de-inlining again in the hopes of making this readable.

- inline the column-based expiration operations as well