sibil/src/lexer/mod.rs

127 lines
3.3 KiB
Rust
Raw Normal View History

2016-12-24 09:05:10 -07:00
/* lexer.rs
* Eryn Wells <eryn@erynwells.me>
*/
pub mod token;
mod char;
mod charset;
mod str;
2016-12-23 17:45:37 -07:00
use self::char::Lexable;
use self::str::CharAt;
use self::str::RelativeIndexable;
use self::token::Token;
use self::token::Kind;
enum State {
Initial,
Identifier,
}
2016-12-20 17:52:29 -08:00
pub struct Lexer {
input: String,
begin: usize,
forward: usize,
state: State,
2016-12-20 17:38:44 -08:00
}
impl Lexer {
2016-12-20 17:52:29 -08:00
pub fn new(input: String) -> Lexer {
Lexer {
input: input,
begin: 0,
forward: 0,
state: State::Initial,
}
}
}
impl Lexer {
fn begin_lexing(&mut self) {
self.forward = self.begin;
self.state = State::Initial;
}
/// Advance the forward pointer to the next character.
fn advance(&mut self) {
2016-12-23 17:46:28 -07:00
self.forward = self.input.index_after(self.forward);
println!("> forward={}", self.forward);
}
/// Retract the forward pointer to the previous character.
fn retract(&mut self) {
2016-12-23 17:46:28 -07:00
self.forward = self.input.index_before(self.forward);
println!("< forward={}", self.forward);
}
2016-12-23 17:53:28 -07:00
/// Advance the begin pointer to prepare for the next iteration.
2016-12-23 17:46:28 -07:00
fn advance_begin(&mut self) {
self.begin = self.input.index_after(self.forward);
println!("> begin={}", self.begin);
}
2016-12-24 09:05:10 -07:00
/// Get the substring between the two input indexes. This is the value to give to a new Token instance.
2016-12-23 17:46:28 -07:00
fn value(&self) -> String {
self.input[self.begin .. self.forward].to_string()
}
}
impl Lexer {
2016-12-23 17:53:28 -07:00
/// Handle self.state == State::Initial
2016-12-23 17:46:28 -07:00
fn state_initial(&mut self, c: char, token: &mut Option<Token>) {
println!("Initial! c='{}'", c);
if c.is_left_paren() {
*token = Some(Token::new(Kind::LeftParen, c.to_string()));
}
else if c.is_right_paren() {
*token = Some(Token::new(Kind::RightParen, c.to_string()));
}
2016-12-24 09:07:38 -07:00
else if c.is_identifier_single() {
*token = Some(Token::new(Kind::Identifier, c.to_string()));
}
2016-12-23 17:46:28 -07:00
else if c.is_identifier_initial() {
self.state = State::Identifier;
self.advance();
}
}
2016-12-23 17:53:28 -07:00
/// Handle self.state == State::Identifier
2016-12-23 17:46:28 -07:00
fn state_identifier(&mut self, c: char, token: &mut Option<Token>) {
if c.is_identifier_subsequent() {
// State in Identifier state.
self.advance();
}
else {
*token = Some(Token::new(Kind::Identifier, self.value()));
self.retract();
}
2016-12-20 17:52:29 -08:00
}
2016-12-20 17:38:44 -08:00
}
impl Iterator for Lexer {
type Item = Token;
fn next(&mut self) -> Option<Token> {
self.begin_lexing();
2016-12-23 17:46:28 -07:00
if self.begin == self.input.len() {
return None;
}
let mut token: Option<Token> = None;
println!("Lexing '{}'", &self.input[self.begin ..]);
2016-12-23 17:53:28 -07:00
while token.is_none() {
2016-12-23 17:46:28 -07:00
if let Some(c) = self.input.char_at(self.forward) {
match self.state {
State::Initial => self.state_initial(c, &mut token),
State::Identifier => self.state_identifier(c, &mut token),
}
}
else {
assert!(false, "Invalid character! :-(");
}
}
2016-12-23 17:46:28 -07:00
self.advance_begin();
2016-12-23 17:53:28 -07:00
assert!(token.is_some(), "We quit the lexing loop but didn't actually have a token. :-(");
2016-12-23 17:46:28 -07:00
token
2016-12-20 17:38:44 -08:00
}
}