NamedDefaults: change in defaulting semantics should be documented in the user's guide

The implementation of GHC proposal #409 has brought with it a change in the way GHC defaults type variables.

GHC is now less conservative: it will allow defaulting type variables, even when they appear as part of non-standard typeclasses. It used to be (as per the Haskell report) that any type variable that appeared in a non-standard constraint would not be defaultable. That restriction has now been lifted (since !11853 (merged) landed, in GHC 9.12), but this has not been documented in the release notes. The NamedDefaults section of the user's guide does explain these new defaulting rules, but not that they apply unconditionally whether NamedDefaults is enabled or not.