More flexibility with persistent type lists?

[Patrick Linskey <pl...@gmail.com>]

Hi,

It seems like a decent number of issues that we see on these lists are
all caused by partial persistent type lists. Perhaps we should be more
permissive with our persistent type discovery by default. For example,
we could auto-detect persistent types related to the listed types
(superclasses and related fields, for example).

Obviously, this would be less spec-compliant. Does it make sense to
have an option to globally control portability vs usability across a
number of aspects of the system, rather than a number of narrower
options?

-Patrick

-- 
Patrick Linskey
202 669 5907