Currently the lexer and parser and using byte offsets to refer to error positions, while the checker is using line numbers, with annotated messages from the source text.
Improve consistency between implementations by making the lexer and parser use the same error handling as the checker.
Currently the lexer and parser and using byte offsets to refer to error positions, while the checker is using line numbers, with annotated messages from the source text.
Improve consistency between implementations by making the lexer and parser use the same error handling as the checker.