What is the state of support for using grammar types and C++? -
Specifically, it would suffice for me if:
1. A type <type> expands to <type>($k) instead of the current union
field selection $k.<type>.
2. Being able to define my own YYSTYPE instead of using the current
union.
I define a root class, which just holds a polymorphic pointer, from
which all the other classes are derived, with no derived class adding
any data. So this would then suffice.
Right now, I need to write out type conversions explicitly, and I do
not use a typed grammar. So having this feature, might be a
simplification.
Hans
