We need to keep track of whitespace tokens in the parse tree or this causes a huge problem for tools as they are not able to preserve user formatting such as preference for line breaks, etc.
In short, input <> parse tree conversion should be physically (not just logically) lossless.
We need to keep track of whitespace tokens in the parse tree or this causes a huge problem for tools as they are not able to preserve user formatting such as preference for line breaks, etc.
In short, input <> parse tree conversion should be physically (not just logically) lossless.