Closed davidweichiang closed 2 years ago
Additionally, @ccshan asks in #71 whether the special cases for handling <>
are really needed. Partly the answer should have been that the special cases eliminate weird things like .0
but apparently I did not do that correctly.
OK, I think the only weird thing is that ()
and <()>
have different types with the same string representation. I plan to try the Haskell way of writing types which will resolve this, but I'm leaving this issue here as a reminder to check whether special cases for <>
can be eliminated.
It all seems to work, but the FGG has nonterminals like
<Unit>.0
, and even theUnit
type has<Unit>.0=()
as its sole inhabitant.