If we allow the right-hand type to be bound first, it can artificially
limit the options for the left-hand side, so attempt to delay binding of
the right-hand side types until after we've chosen something for the
left-hand side.
The test here will not fail without this change, but will fail without
this change if the ConstraintGraph.cpp changes from @xedin's edge
contraction patch (https://github.com/apple/swift/pull/11118) are
applied.
When search for potential bindings for the given type variable
encounters `DynamicTypeOf` constraint, its validation should use
representative type variable of the left-hand side instead of
direct comparison.
When trying to find bindings let's avoid type variables which
represent left-hand side of the `DynamicTypeOf` constraint,
because such type variables might be (implicitly) convertible to
[existential] metatype, but at the same time bound to something else
(e.g. Any), which fails when solver tries to simplify `DynamicTypeOf`
constraint itself.
When we saw lvalue- or inout-types we were not appropriately looking at
the underlying types when performing a join. As a result we would infer
Any in cases where we should have inferred a common supertype.
rdar://problem/31127419
Split out the code for selecting potential bindings into a separate file
as a first step before refactoring it for improved clarity and ease of
modification.
