Open yutakang opened 6 years ago
It turned out to be a bad idea to convert un-curried syntax trees to tables. And I found out that it is more intuitive, straightforward, and functional to work directly on un-curried syntax trees.
However, tagging each node in un-curried syntax tree with a unique identifier still might be a good idea when handling induct_tac
and de-Bruijn indices in future.
Currently, each
point
has only three fields:cname
,utyp
, andlevel
. Therefore, it is possible that a syntax tree has multiplepoint
s of the same value. We should avoid such situation.