I just wanted to drop a line here saying "it will take me another couple of
days to deal with this".
On one hand, I want to wait with this for Marc's work to land. On the other,
there are more urgent issues even in Oracle-land (the broken Py3 tests).
And yes, I'm with Anssi on removing get_db_prep_lookup() -- the partition
between "make ready for field type" and "make ready for lookup type" that is in
the current code is weird and causes problems, which custom lookups solve.
However, note that if Anssi is correct about the backwards-compatibility
point, then Marc's work changing Contains may not be backwards-compatible.
More later,
Shai.
