pub fn tokenize_with_offsets(
input: &str,
backslash_escapes: bool,
) -> Result<(Vec<Token>, Vec<usize>), LexError>Expand description
v7.39 (read01 round 95) — like tokenize_with but also returns, for each
token, the byte offset in input where it started (the Eof token maps to
input.len()). The parser uses this to translate a failing token index into
PG’s 1-based character error position (the ErrorResponse P field that psql
renders as LINE n: … ^).