Continue work on parser, improved lexer

- Added literal parsing (improved lexing)
- Revised error handling
This commit is contained in:
Erin 2021-04-27 10:51:39 +02:00 committed by ondra05
parent cc4ec803c4
commit 3b8ce34c2b
5 changed files with 92 additions and 63 deletions

View file

@ -1,10 +1,2 @@
functio test() {
functio nested() {
var c = false;
}
var a = true;
}
functio another() {
var b = false;
}
var a = 3;
var b = 4;

View file

@ -1,6 +1,18 @@
use crate::variables::Value;
#[derive(Debug, Clone)]
pub enum Expr {
VariableDeclaration { iden: String, init: Option<String> },
FunctionDeclaration { iden: String, body: Vec<Expr> },
BfFDeclaration { iden: String, code: String },
VariableDeclaration {
iden: String,
init: Option<Box<Expr>>,
},
FunctionDeclaration {
iden: String,
body: Vec<Expr>,
},
BfFDeclaration {
iden: String,
code: String,
},
Literal(Value),
}

View file

@ -3,8 +3,11 @@ mod utils;
use item::Expr;
use crate::error::{Error, ErrorKind};
use crate::tokens::Token;
use crate::{
error::{Error, ErrorKind},
variables::Value,
};
use logos::Logos;
@ -29,12 +32,13 @@ impl<'a> Parser<'a> {
if token.is_none() {
return Ok(self.ast.clone());
};
let expr = self.parse_expr(&token)?;
let expr = self.parse_expr(token)?;
self.ast.push(expr);
}
}
fn parse_expr(&mut self, token: &Option<Token>) -> Result<Expr, Error> {
fn parse_expr(&mut self, token: Option<Token>) -> Result<Expr, Error> {
if matches!(token, None) {
return Err(Error {
kind: ErrorKind::EndOfTokenStream,
@ -42,36 +46,23 @@ impl<'a> Parser<'a> {
});
}
Ok(todo!())
}
let token = token.unwrap();
let start = self.lexer.span().start;
/*
/// Start parsing Token Vector into Abstract Syntax Tree
pub fn parse(&mut self) -> Vec<Expr> {
let mut ast = vec![];
while let Some(token) = self.lexer.next() {
let expr = match token {
Token::Variable => self.variable_declaration(),
Token::Function => self.function_declaration(),
Token::BfFunction => self.bff_declaration(),
Token::RightBrace => return ast,
_ => Err(Error {
kind: ErrorKind::SyntaxError,
position: 0..0,
}),
};
match expr {
Ok(o) => ast.push(o),
Err(e) => {
e.panic(self.lexer.slice());
break;
}
}
match token {
Token::Variable => self.variable_declaration(),
Token::Function => self.function_declaration(),
Token::BfFunction => self.bff_declaration(),
Token::String(x) => Ok(Expr::Literal(Value::Str(x))),
Token::Integer(x) => Ok(Expr::Literal(Value::Int(x))),
Token::Boolean(x) => Ok(Expr::Literal(Value::Bool(x))),
Token::Aboolean(x) => Ok(Expr::Literal(Value::Abool(x))),
_ => Err(Error {
kind: ErrorKind::SyntaxError("Unexpected identifier".to_owned()),
position: start..self.lexer.span().end,
}),
}
ast
}
*/
/// Parse variable declaration
///
@ -82,9 +73,10 @@ impl<'a> Parser<'a> {
let init = match self.lexer.next() {
Some(Token::Semicolon) => None,
Some(Token::Assignment) => {
let value = self.require(Token::Boolean)?; // TODO: Shouldn't be limited to boolean (pattern match?)
let value = self.lexer.next();
let value = self.parse_expr(value)?; // TODO: Shouldn't be limited to boolean (pattern match?)
self.require(Token::Semicolon)?;
Some(value)
Some(Box::new(value))
}
_ => {
return Err(Error {
@ -106,7 +98,9 @@ impl<'a> Parser<'a> {
// TODO: Arguments
self.require(Token::RightParenthesis)?;
self.require(Token::LeftBrace)?;
let body = vec![];
let expr = self.lexer.next();
let expr = self.parse_expr(expr);
let body = vec![expr?];
Ok(Expr::FunctionDeclaration { iden, body })
}
@ -115,6 +109,7 @@ impl<'a> Parser<'a> {
///
/// `bff [iden] { ... }`
fn bff_declaration(&mut self) -> Result<Expr, Error> {
// TODO: Make it throw error when EOF
let iden = self.require(Token::Identifier)?;
self.require(Token::LeftBrace)?;
let mut code = String::new();
@ -129,10 +124,9 @@ impl<'a> Parser<'a> {
| Token::LeftBracket
| Token::RightBracket => self.lexer.slice(),
Token::RightBrace => break,
_ => break,
_ => return Err(self.unexpected_token(None)),
});
}
self.require(Token::RightBrace)?;
Ok(Expr::BfFDeclaration { iden, code })
}
}

View file

@ -23,13 +23,21 @@ pub fn num2abool(number: i32) -> Abool {
impl<'a> Parser<'a> {
/// Require type of token as next and return it's value (sometimes irrelevant)
pub(super) fn require(&mut self, with: Token) -> Result<String, Error> {
if self.lexer.next() == Some(with) {
if self.lexer.next() == Some(with.clone()) {
Ok(self.lexer.slice().to_owned())
} else {
Err(Error {
kind: ErrorKind::SyntaxError("Mysterious parse error".to_owned()),
position: self.lexer.span(),
})
Err(self.unexpected_token(Some(with)))
}
}
pub(super) fn unexpected_token(&mut self, expected: Option<Token>) -> Error {
Error {
kind: ErrorKind::SyntaxError(format!(
"Unexpected token: `{}` (required: `{:?}`)",
self.lexer.slice(),
expected
)),
position: self.lexer.span(),
}
}
}

View file

@ -1,23 +1,25 @@
use logos::Logos;
use logos::{Lexer, Logos};
#[derive(Logos, Debug, PartialEq)]
use crate::variables::Abool;
#[derive(Logos, Debug, PartialEq, Clone)]
pub enum Token {
// Literals
/// True, False
#[regex("true|false")]
Boolean,
#[regex("true|false", get_bool)]
Boolean(bool),
/// Always, Sometimes, Never
#[regex("always|sometimes|never")]
Aboolean,
#[regex("always|sometimes|never", get_abool)]
Aboolean(Abool),
/// String
#[regex("\"(\\.|[^\"])*\"")]
String,
#[regex("\"(\\.|[^\"])*\"", get_string)]
String(String),
/// Integer
#[regex(r"[0-9]+")]
Integer,
#[regex(r"[0-9]+", get_int)]
Integer(i32),
/// A C-complaint identifier
#[regex(r"[a-zA-Z_][a-zA-Z_0-9]*")]
@ -122,3 +124,24 @@ pub enum Token {
#[error]
Error,
}
fn get_bool(lexer: &mut Lexer<Token>) -> Option<bool> {
lexer.slice().parse().ok()
}
fn get_int(lexer: &mut Lexer<Token>) -> Option<i32> {
lexer.slice().parse().ok()
}
fn get_string(lexer: &mut Lexer<Token>) -> String {
lexer.slice().to_owned()
}
fn get_abool(lexer: &mut Lexer<Token>) -> Option<Abool> {
match lexer.slice() {
"always" => Some(Abool::Always),
"sometimes" => Some(Abool::Sometimes),
"never" => Some(Abool::Never),
_ => None,
}
}