I was noticing that the reported column position for identifiers points to the whitespace before the identifier, not the identifier itself. I think it's in the S-expression reader, not the parser code, because it's wrong on the raw SemgusToken objects.
I was noticing that the reported column position for identifiers points to the whitespace before the identifier, not the identifier itself. I think it's in the S-expression reader, not the parser code, because it's wrong on the raw
SemgusToken
objects.