rust/src/lexer/mod.rs

108 lines
2.7 KiB
Rust
Raw Normal View History

2017-12-29 14:33:04 -06:00
use {Token, SyntaxKind};
use syntax_kinds::*;
mod ptr;
use self::ptr::Ptr;
2017-12-29 15:48:47 -06:00
mod classes;
use self::classes::*;
2017-12-30 06:56:52 -06:00
mod numbers;
use self::numbers::scan_number;
2017-12-29 14:33:04 -06:00
pub fn next_token(text: &str) -> Token {
assert!(!text.is_empty());
let mut ptr = Ptr::new(text);
let c = ptr.bump().unwrap();
let kind = next_token_inner(c, &mut ptr);
let len = ptr.into_len();
Token { kind, len }
}
fn next_token_inner(c: char, ptr: &mut Ptr) -> SyntaxKind {
// Note: r as in r" or r#" is part of a raw string literal,
// b as in b' is part of a byte literal.
// They are not identifiers, and are handled further down.
2017-12-29 15:48:47 -06:00
let ident_start = is_ident_start(c) && !string_literal_start(c, ptr.next(), ptr.nnext());
2017-12-29 14:33:04 -06:00
if ident_start {
2017-12-30 06:22:40 -06:00
return scan_ident(c, ptr);
2017-12-29 14:33:04 -06:00
}
2017-12-29 15:48:47 -06:00
if is_whitespace(c) {
ptr.bump_while(is_whitespace);
return WHITESPACE;
}
2017-12-29 14:33:04 -06:00
2017-12-30 06:22:40 -06:00
if is_dec_digit(c) {
return scan_number(c, ptr);
}
2017-12-30 07:30:37 -06:00
// One-byte tokens.
match c {
';' => return SEMI,
',' => return COMMA,
'(' => return L_PAREN,
')' => return R_PAREN,
'{' => return L_CURLY,
'}' => return R_CURLY,
'[' => return L_BRACK,
']' => return R_BRACK,
'@' => return AT,
'#' => return POUND,
'~' => return TILDE,
'?' => return QUESTION,
'$' => return DOLLAR,
'.' => return match (ptr.next(), ptr.nnext()) {
(Some('.'), Some('.')) => {
ptr.bump();
ptr.bump();
DOTDOTDOT
},
(Some('.'), Some('=')) => {
ptr.bump();
ptr.bump();
DOTDOTEQ
},
(Some('.'), _) => {
ptr.bump();
DOTDOT
},
_ => DOT
},
':' => return match ptr.next() {
Some(':') => {
ptr.bump();
COLONCOLON
}
_ => COLON
},
_ => (),
}
2017-12-29 15:53:06 -06:00
ERROR
2017-12-29 14:33:04 -06:00
}
2017-12-30 06:22:40 -06:00
fn scan_ident(c: char, ptr: &mut Ptr) -> SyntaxKind {
let is_single_letter = match ptr.next() {
None => true,
Some(c) if !is_ident_continue(c) => true,
_ => false,
};
if is_single_letter {
return if c == '_' { UNDERSCORE } else { IDENT };
}
ptr.bump_while(is_ident_continue);
IDENT
}
2017-12-29 14:33:04 -06:00
fn string_literal_start(c: char, c1: Option<char>, c2: Option<char>) -> bool {
match (c, c1, c2) {
('r', Some('"'), _) |
('r', Some('#'), _) |
('b', Some('"'), _) |
('b', Some('\''), _) |
('b', Some('r'), Some('"')) |
('b', Some('r'), Some('#')) => true,
_ => false
}
}