Suggestion for better errors

[Quelklef]

When parsing, include a field range which is the range of text that the node takes.

For example, in

main: "banana"

the node for main would have the range [0, 13], from ma to a".

This would allow for much better error messages that come after parsing.