Stupid redundancy in Nim's semmed AST #25302
Description
Refactoring: If prc.typ.n contains the semchecked parameters and n.ast[paramPos] does not, it should not contain misleading nodes. Either make it contain nkEmpty or find a way to make prc.typ.n a thing of the past.
Speaking of which the fact that we fill both t.sons AND t.n with the parameters is wasteful.