Compare commits
9 Commits
613e839a6e
...
master
Author | SHA1 | Date | |
---|---|---|---|
d377c8448f | |||
af30410aab | |||
360687f3c0 | |||
f7ec10646b | |||
ad57170010 | |||
1a2563f756 | |||
d70b196042 | |||
771363106f | |||
384f382cfc |
13
flake.nix
13
flake.nix
@@ -9,25 +9,28 @@
|
||||
inputs.nixpkgs.follows = "nixpkgs";
|
||||
};
|
||||
|
||||
rust-overlay = {
|
||||
url = github:oxalica/rust-overlay;
|
||||
fenix = {
|
||||
url = github:nix-community/fenix;
|
||||
inputs.nixpkgs.follows = "nixpkgs";
|
||||
};
|
||||
|
||||
flake-utils.url = github:numtide/flake-utils;
|
||||
};
|
||||
|
||||
outputs = inputs@{ self, nixpkgs, crane, rust-overlay, flake-utils }:
|
||||
outputs = inputs@{ self, nixpkgs, crane, fenix, flake-utils }:
|
||||
flake-utils.lib.eachDefaultSystem (system:
|
||||
let
|
||||
pkgs = import nixpkgs {
|
||||
inherit system;
|
||||
overlays = [ rust-overlay.overlays.default ];
|
||||
overlays = [ fenix.overlays.default ];
|
||||
};
|
||||
|
||||
inherit (pkgs) lib;
|
||||
|
||||
toolchain = pkgs.rust-bin.fromRustupToolchainFile ./rust-toolchain.toml;
|
||||
toolchain = pkgs.fenix.fromToolchainFile {
|
||||
file = ./rust-toolchain.toml;
|
||||
sha256 = "sha256-n8LtGbpj/yCUGo0NFJ7FNv9fSdT9oKEUl+EPLg06JdQ=";
|
||||
};
|
||||
|
||||
craneLib = (crane.mkLib pkgs).overrideToolchain toolchain;
|
||||
src = craneLib.cleanCargoSource (craneLib.path ./.);
|
||||
|
@@ -1,4 +1,5 @@
|
||||
use std::env;
|
||||
use std::path::PathBuf;
|
||||
use std::process::exit;
|
||||
|
||||
const VERSION: &str = env!("CARGO_PKG_VERSION");
|
||||
@@ -8,7 +9,7 @@ const CRATE: &str = env!("CARGO_CRATE_NAME");
|
||||
#[derive(Default)]
|
||||
pub struct Args {
|
||||
version: bool,
|
||||
file: Option<String>,
|
||||
file: Option<PathBuf>,
|
||||
}
|
||||
|
||||
impl Args {
|
||||
@@ -34,7 +35,7 @@ impl Args {
|
||||
if self.file.is_some() {
|
||||
panic!("please specify only a single source file!");
|
||||
}
|
||||
self.file = Some(file.to_owned());
|
||||
self.file = Some(PathBuf::from(file));
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -55,7 +56,7 @@ impl Args {
|
||||
/// Fetches the file from the arguments.
|
||||
/// Panics if there is no file in the arguments
|
||||
#[inline]
|
||||
pub fn get_file(self) -> String {
|
||||
pub fn get_file(self) -> PathBuf {
|
||||
self.file.expect("no file supplied!")
|
||||
}
|
||||
}
|
||||
|
113
src/ast.rs
113
src/ast.rs
@@ -1,89 +1,132 @@
|
||||
/// A very naive AST definition using recursive enums
|
||||
/// See the parser for implementation
|
||||
//! A very naive AST definition using recursive enums
|
||||
//!
|
||||
//! See the parser for implementation
|
||||
|
||||
use std::rc::Rc;
|
||||
|
||||
pub type Parent = Vec<Entity>;
|
||||
|
||||
/// Entities are functions, classes, and modules
|
||||
#[derive(Debug)]
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum Entity {
|
||||
Fn(Fn),
|
||||
Class(Class),
|
||||
Module(Module),
|
||||
Static(Let),
|
||||
}
|
||||
|
||||
#[derive(Debug)]
|
||||
/// A module just provides an additional scope
|
||||
///
|
||||
/// TODO: Add exporting and importing modules
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub struct Module {
|
||||
/// Name of module
|
||||
pub name: Rc<str>,
|
||||
/// Everything inside the module
|
||||
pub children: Vec<ModuleChildren>,
|
||||
}
|
||||
|
||||
/// Modules contain functions, classes and statements
|
||||
#[derive(Debug)]
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum ModuleChildren {
|
||||
Fn(Fn),
|
||||
Class(Class),
|
||||
Statement(Statement),
|
||||
Module(Module),
|
||||
Static(Let),
|
||||
}
|
||||
|
||||
#[derive(Debug)]
|
||||
/// Classes encapsulate functions and definitions.
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub struct Class {
|
||||
/// Name of class
|
||||
pub name: Rc<str>,
|
||||
/// Everything inside the class
|
||||
pub children: Vec<ClassChildren>,
|
||||
}
|
||||
|
||||
/// Classes contain functions and statements.
|
||||
///
|
||||
/// TODO: Maybe change statements to something else
|
||||
#[derive(Debug)]
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum ClassChildren {
|
||||
Fn(Fn),
|
||||
Statement(Statement),
|
||||
Let(Let),
|
||||
Static(Let),
|
||||
}
|
||||
|
||||
#[derive(Debug)]
|
||||
/// A Function
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub struct Fn {
|
||||
/// Name of the function
|
||||
pub name: Rc<str>,
|
||||
pub return_typ: Option<Primitive>,
|
||||
pub params: Vec<(Rc<str>, Primitive)>,
|
||||
/// Optional return type
|
||||
pub return_ty: Option<Ty>,
|
||||
/// Parameters
|
||||
pub params: Vec<(Rc<str>, Ty)>,
|
||||
/// The function block
|
||||
pub children: Vec<Statement>,
|
||||
}
|
||||
|
||||
#[derive(Debug)]
|
||||
/// Statements encapsulate expressions and definitions
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum Statement {
|
||||
Static(Let),
|
||||
Let(Let),
|
||||
Expr(Expr),
|
||||
Block(Vec<Statement>),
|
||||
}
|
||||
|
||||
#[derive(Debug)]
|
||||
/// A variable definition
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub struct Let {
|
||||
/// Name of variabe
|
||||
pub name: Rc<str>,
|
||||
pub typ: Primitive,
|
||||
/// Type of variable
|
||||
pub ty: Ty,
|
||||
/// Value of variable
|
||||
pub expr: Option<Expr>,
|
||||
}
|
||||
|
||||
type Op = crate::lexer::TokenSymbol;
|
||||
|
||||
#[derive(Debug)]
|
||||
pub enum Expr {
|
||||
Int(i32),
|
||||
Float(f32),
|
||||
Char(char),
|
||||
Op(Op, Box<Expr>, Option<Box<Expr>>),
|
||||
If(Box<Expr>, Box<Expr>, Option<Box<Expr>>),
|
||||
Loop,
|
||||
Break,
|
||||
Continue,
|
||||
}
|
||||
|
||||
/// Primitives
|
||||
///
|
||||
/// TODO: add arrays and pointers maybe
|
||||
#[derive(Debug)]
|
||||
pub enum Primitive {
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum Ty {
|
||||
Int,
|
||||
Float,
|
||||
Char,
|
||||
}
|
||||
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub struct If {
|
||||
pub cond: Box<Expr>,
|
||||
pub then: Vec<Statement>,
|
||||
pub or: Option<Box<ElseType>>,
|
||||
}
|
||||
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum ElseType {
|
||||
If(If),
|
||||
Else(Vec<Statement>),
|
||||
}
|
||||
|
||||
pub(crate) type Op = crate::lexer::TokenSymbol;
|
||||
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum Literal {
|
||||
Int(i32),
|
||||
Float(f32),
|
||||
Char(char),
|
||||
}
|
||||
|
||||
/// Lowest form of expression
|
||||
///
|
||||
/// TODO: refine
|
||||
#[derive(Debug, PartialEq)]
|
||||
pub enum Expr {
|
||||
Literal(Literal),
|
||||
Identifier(Rc<str>),
|
||||
Op(Op, Box<Expr>, Option<Box<Expr>>),
|
||||
If(If),
|
||||
Block(Vec<Statement>),
|
||||
Loop(Vec<Statement>),
|
||||
Break,
|
||||
Continue,
|
||||
Return(Option<Box<Expr>>),
|
||||
}
|
||||
|
90
src/lexer.rs
90
src/lexer.rs
@@ -6,7 +6,7 @@ use std::str;
|
||||
/// All token literals
|
||||
///
|
||||
/// TODO: Add string
|
||||
#[derive(Debug, PartialEq)]
|
||||
#[derive(Debug, PartialEq, Clone, Copy)]
|
||||
pub enum TokenLiteral {
|
||||
Int,
|
||||
Float,
|
||||
@@ -14,11 +14,9 @@ pub enum TokenLiteral {
|
||||
}
|
||||
|
||||
/// All token symbols
|
||||
///
|
||||
/// TODO: Maybe add *
|
||||
#[derive(Debug, PartialEq)]
|
||||
#[derive(Debug, PartialEq, Clone, Copy)]
|
||||
pub enum TokenSymbol {
|
||||
// operators
|
||||
// arithmetic
|
||||
Plus,
|
||||
Minus,
|
||||
Star,
|
||||
@@ -40,9 +38,7 @@ pub enum TokenSymbol {
|
||||
|
||||
// relational
|
||||
Gt,
|
||||
Ge,
|
||||
Lt,
|
||||
Le,
|
||||
GtEq,
|
||||
LtEq,
|
||||
EqEq,
|
||||
@@ -69,7 +65,7 @@ pub enum TokenSymbol {
|
||||
}
|
||||
|
||||
/// All token keywod
|
||||
#[derive(Debug, PartialEq)]
|
||||
#[derive(Debug, PartialEq, Clone, Copy)]
|
||||
pub enum TokenKeyword {
|
||||
// parents
|
||||
Fn,
|
||||
@@ -77,6 +73,7 @@ pub enum TokenKeyword {
|
||||
Module,
|
||||
|
||||
// statements
|
||||
Static,
|
||||
Let,
|
||||
Ret,
|
||||
|
||||
@@ -89,6 +86,7 @@ pub enum TokenKeyword {
|
||||
Loop,
|
||||
Break,
|
||||
Continue,
|
||||
Return,
|
||||
|
||||
// primitives
|
||||
Int,
|
||||
@@ -99,7 +97,7 @@ pub enum TokenKeyword {
|
||||
/// All token delimiters
|
||||
///
|
||||
/// TODO: Maybe add \[ and \]
|
||||
#[derive(Debug, PartialEq)]
|
||||
#[derive(Debug, PartialEq, Clone, Copy)]
|
||||
pub enum TokenDelimiter {
|
||||
BraceOpen,
|
||||
BraceClose,
|
||||
@@ -108,7 +106,7 @@ pub enum TokenDelimiter {
|
||||
}
|
||||
|
||||
/// All tokens
|
||||
#[derive(Debug, PartialEq)]
|
||||
#[derive(Debug, PartialEq, Clone, Copy)]
|
||||
pub enum TokenKind {
|
||||
Newline,
|
||||
Eof,
|
||||
@@ -138,9 +136,10 @@ pub struct Lexer<'a> {
|
||||
/// A peekable double ended queue for the tokens
|
||||
tokens: VecDeque<Token>,
|
||||
/// Current line number
|
||||
pub line: usize,
|
||||
pub(crate) line: usize,
|
||||
pub(crate) col: usize,
|
||||
/// Start character index for the current token
|
||||
pub start: usize,
|
||||
start: usize,
|
||||
/// End character index for the current token
|
||||
end: usize,
|
||||
}
|
||||
@@ -172,6 +171,7 @@ impl<'a> Lexer<'a> {
|
||||
chars: content.chars().peekable(),
|
||||
tokens: VecDeque::new(),
|
||||
line: 1,
|
||||
col: 1,
|
||||
start: 0,
|
||||
end: 0,
|
||||
}
|
||||
@@ -198,6 +198,7 @@ impl<'a> Lexer<'a> {
|
||||
#[inline]
|
||||
fn next(&mut self) -> Option<char> {
|
||||
self.end += 1;
|
||||
self.col += 1;
|
||||
self.chars.next()
|
||||
}
|
||||
|
||||
@@ -233,11 +234,16 @@ impl<'a> Lexer<'a> {
|
||||
'0'..='9' => {}
|
||||
'.' => {
|
||||
if is_float {
|
||||
self.error("Multiple decimals encountered");
|
||||
self.error("multiple decimals encountered");
|
||||
return self.new_token(TokenKind::Invalid);
|
||||
}
|
||||
is_float = true;
|
||||
}
|
||||
'e' | 'E' => {
|
||||
self.next();
|
||||
is_float = true;
|
||||
break;
|
||||
}
|
||||
_ => break,
|
||||
}
|
||||
self.next();
|
||||
@@ -259,6 +265,11 @@ impl<'a> Lexer<'a> {
|
||||
return self.new_token(TokenKind::Invalid);
|
||||
}
|
||||
|
||||
if self.peek() != Some(&'\'') {
|
||||
self.error("Expected '");
|
||||
return self.new_token(TokenKind::Invalid);
|
||||
}
|
||||
|
||||
// skip '
|
||||
self.next();
|
||||
|
||||
@@ -268,7 +279,7 @@ impl<'a> Lexer<'a> {
|
||||
fn get_alphanumeric(&mut self) -> Token {
|
||||
while let Some(c) = self.peek() {
|
||||
match c {
|
||||
'a'..='z' | 'A'..='Z' | '0'..='9' => {}
|
||||
'a'..='z' | 'A'..='Z' | '0'..='9' | '_' => {}
|
||||
_ => break,
|
||||
}
|
||||
self.next();
|
||||
@@ -281,6 +292,7 @@ impl<'a> Lexer<'a> {
|
||||
"fn" => Keyword(Fn),
|
||||
"class" => Keyword(Class),
|
||||
"module" => Keyword(Module),
|
||||
"static" => Keyword(Static),
|
||||
"let" => Keyword(Let),
|
||||
"ret" => Keyword(Ret),
|
||||
"if" => Keyword(If),
|
||||
@@ -289,6 +301,7 @@ impl<'a> Lexer<'a> {
|
||||
"loop" => Keyword(Loop),
|
||||
"break" => Keyword(Break),
|
||||
"continue" => Keyword(Continue),
|
||||
"return" => Keyword(Return),
|
||||
"int" => Keyword(Int),
|
||||
"float" => Keyword(Float),
|
||||
"char" => Keyword(Char),
|
||||
@@ -392,10 +405,11 @@ impl<'a> Lexer<'a> {
|
||||
'\n' => {
|
||||
self.next();
|
||||
self.line += 1;
|
||||
self.col = 0;
|
||||
self.new_token(TokenKind::Newline)
|
||||
}
|
||||
'0'..='9' => self.get_numeric(),
|
||||
'a'..='z' | 'A'..='Z' => self.get_alphanumeric(),
|
||||
'a'..='z' | 'A'..='Z' | '_' => self.get_alphanumeric(),
|
||||
'\'' => self.get_char(),
|
||||
_ => self.get_symbol(),
|
||||
}
|
||||
@@ -441,7 +455,7 @@ fn test_peek_next() {
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_tokens_1() {
|
||||
fn test_tokens() {
|
||||
let mut lexer = Lexer::new("let test02 = 4 << 1");
|
||||
|
||||
use TokenKind::*;
|
||||
@@ -501,9 +515,51 @@ fn test_tokens_2() {
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_tokens_3() {
|
||||
fn test_tokens_eof() {
|
||||
let mut lexer = Lexer::new("");
|
||||
|
||||
assert_eq!(lexer.peek_token().kind, TokenKind::Eof);
|
||||
assert_eq!(lexer.next_token().kind, TokenKind::Eof);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_tokens_numeric() {
|
||||
let mut lexer = Lexer::new("3342");
|
||||
|
||||
let token = lexer.next_token();
|
||||
assert_eq!(token.kind, TokenKind::Literal(TokenLiteral::Int));
|
||||
assert_eq!(*token.val, *"3342");
|
||||
|
||||
assert_eq!(lexer.next_token().kind, TokenKind::Eof);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_tokens_numeric_2() {
|
||||
let mut lexer = Lexer::new("334.2e");
|
||||
|
||||
let token = lexer.next_token();
|
||||
assert_eq!(token.kind, TokenKind::Literal(TokenLiteral::Float));
|
||||
assert_eq!(*token.val, *"334.2e");
|
||||
|
||||
assert_eq!(lexer.next_token().kind, TokenKind::Eof);
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_tokens_numeric_3() {
|
||||
let mut lexer = Lexer::new("334.2e-5");
|
||||
|
||||
let mut token = lexer.next_token();
|
||||
assert_eq!(token.kind, TokenKind::Literal(TokenLiteral::Float));
|
||||
assert_eq!(*token.val, *"334.2e");
|
||||
|
||||
assert_eq!(
|
||||
lexer.next_token().kind,
|
||||
TokenKind::Symbol(TokenSymbol::Minus)
|
||||
);
|
||||
|
||||
token = lexer.next_token();
|
||||
assert_eq!(token.kind, TokenKind::Literal(TokenLiteral::Int));
|
||||
assert_eq!(*token.val, *"5");
|
||||
|
||||
assert_eq!(lexer.next_token().kind, TokenKind::Eof);
|
||||
}
|
||||
|
@@ -1,3 +1,4 @@
|
||||
pub mod args;
|
||||
pub mod ast;
|
||||
pub mod lexer;
|
||||
pub mod parser;
|
||||
|
20
src/main.rs
20
src/main.rs
@@ -4,10 +4,7 @@ use std::{
|
||||
};
|
||||
|
||||
use tricc::args::Args;
|
||||
use tricc::lexer::{
|
||||
Lexer,
|
||||
TokenKind,
|
||||
};
|
||||
use tricc::parser::Parser;
|
||||
|
||||
fn main() {
|
||||
panic::set_hook(Box::new(|panic_info| {
|
||||
@@ -31,10 +28,13 @@ fn main() {
|
||||
|
||||
let file = args.get_file();
|
||||
let content = fs::read_to_string(&file).expect("Couldn't read the file");
|
||||
|
||||
let mut lexer = Lexer::new(content.as_str());
|
||||
|
||||
while lexer.peek_token().kind != TokenKind::Eof {
|
||||
println!("{:?}", lexer.next_token());
|
||||
}
|
||||
let mut parser = Parser::new(&content);
|
||||
let Some(parent) = parser.parse() else {
|
||||
eprintln!(
|
||||
"Failed to parse {} - See the errors above",
|
||||
file.to_string_lossy()
|
||||
);
|
||||
std::process::exit(1);
|
||||
};
|
||||
println!("Parsed AST:\n{:#?}", parent);
|
||||
}
|
||||
|
215
src/parser/entity.rs
Normal file
215
src/parser/entity.rs
Normal file
@@ -0,0 +1,215 @@
|
||||
use super::Parser;
|
||||
use crate::ast::*;
|
||||
use crate::lexer::{
|
||||
TokenDelimiter,
|
||||
TokenKeyword,
|
||||
TokenKind,
|
||||
TokenSymbol,
|
||||
};
|
||||
use std::rc::Rc;
|
||||
|
||||
impl<'a> Parser<'a> {
|
||||
/// entity ::= module | class | fn | static
|
||||
pub(super) fn parse_entity(&mut self) -> Option<Entity> {
|
||||
use TokenKeyword::*;
|
||||
let token = self.peek_token();
|
||||
|
||||
if let TokenKind::Keyword(keyword) = &token.kind {
|
||||
Some(match keyword {
|
||||
Module => Entity::Module(self.parse_module()?),
|
||||
Class => Entity::Class(self.parse_class()?),
|
||||
Fn => Entity::Fn(self.parse_fn()?),
|
||||
Static => Entity::Static(self.parse_static()?),
|
||||
_ => {
|
||||
self.error_expected_peek("entity");
|
||||
return None;
|
||||
}
|
||||
})
|
||||
} else {
|
||||
self.error_expected_peek("entity");
|
||||
None
|
||||
}
|
||||
}
|
||||
|
||||
/// module ::= "module" ident "{" { module | fn | static | class } "}"
|
||||
fn parse_module(&mut self) -> Option<Module> {
|
||||
self.next_token();
|
||||
|
||||
let name = self.parse_ident()?;
|
||||
let mut children = vec![];
|
||||
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceOpen)) {
|
||||
self.error_expected_peek("{");
|
||||
return None;
|
||||
}
|
||||
|
||||
loop {
|
||||
use TokenKeyword::*;
|
||||
self.trim_newlines();
|
||||
if let TokenKind::Keyword(keyword) = &self.peek_token().kind {
|
||||
children.push(match keyword {
|
||||
Module => ModuleChildren::Module(self.parse_module()?),
|
||||
Fn => ModuleChildren::Fn(self.parse_fn()?),
|
||||
Static => ModuleChildren::Static(self.parse_static()?),
|
||||
Class => ModuleChildren::Class(self.parse_class()?),
|
||||
_ => {
|
||||
self.error_expected_peek("module child");
|
||||
return None;
|
||||
}
|
||||
});
|
||||
if !self.check_newline_or_tok(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("newline or }");
|
||||
return None;
|
||||
}
|
||||
} else if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("}");
|
||||
return None;
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
Some(Module { name, children })
|
||||
}
|
||||
|
||||
/// class ::= "class" ident "{" { fn | static | let } "}"
|
||||
fn parse_class(&mut self) -> Option<Class> {
|
||||
self.next_token();
|
||||
|
||||
let name = self.parse_ident()?;
|
||||
let mut children = vec![];
|
||||
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceOpen)) {
|
||||
self.error_expected_peek("{");
|
||||
return None;
|
||||
}
|
||||
|
||||
loop {
|
||||
use TokenKeyword::*;
|
||||
self.trim_newlines();
|
||||
if let TokenKind::Keyword(keyword) = &self.peek_token().kind {
|
||||
children.push(match keyword {
|
||||
Fn => ClassChildren::Fn(self.parse_fn()?),
|
||||
Static => ClassChildren::Static(self.parse_static()?),
|
||||
Let => ClassChildren::Let(self.parse_let()?),
|
||||
_ => {
|
||||
self.error_expected_peek("class child");
|
||||
return None;
|
||||
}
|
||||
});
|
||||
if !self.check_newline_or_tok(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("newline or }");
|
||||
return None;
|
||||
}
|
||||
} else if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("}");
|
||||
return None;
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
|
||||
Some(Class { name, children })
|
||||
}
|
||||
|
||||
/// fn ::= "fn" ident "(" [ identWithTy { "," identWithTy } ] ")" [ ":" ty ]
|
||||
/// "{" { statement } "}"
|
||||
fn parse_fn(&mut self) -> Option<Fn> {
|
||||
self.next_token();
|
||||
|
||||
let name = self.parse_ident()?;
|
||||
let mut params: Vec<(Rc<str>, Ty)> = vec![];
|
||||
let mut return_typ: Option<Ty> = None;
|
||||
let mut children: Vec<Statement> = vec![];
|
||||
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::ParenOpen)) {
|
||||
self.error_expected_peek("(");
|
||||
return None;
|
||||
}
|
||||
|
||||
loop {
|
||||
if self.peek_token().kind == TokenKind::Identifier {
|
||||
params.push(self.parse_ident_with_ty()?);
|
||||
}
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(TokenSymbol::Comma)) {
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::ParenClose)) {
|
||||
self.error_expected_peek(", or )");
|
||||
return None;
|
||||
} else {
|
||||
break;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
if self.skip_token(TokenKind::Symbol(TokenSymbol::Colon)) {
|
||||
return_typ = Some(self.parse_ty()?);
|
||||
}
|
||||
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceOpen)) {
|
||||
self.error_expected_peek("{");
|
||||
return None;
|
||||
}
|
||||
|
||||
loop {
|
||||
self.trim_newlines();
|
||||
if self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
break;
|
||||
}
|
||||
children.push(self.parse_statement()?);
|
||||
if !self.check_newline_or_tok(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("newline or }");
|
||||
return None;
|
||||
}
|
||||
}
|
||||
|
||||
Some(Fn {
|
||||
name,
|
||||
return_ty: return_typ,
|
||||
params,
|
||||
children,
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_parse_entity() {
|
||||
let mut parser = Parser::new(
|
||||
r#"module module01 {
|
||||
class class01 {
|
||||
fn fn01(param01: char, param02: float) {
|
||||
static let let01: int = 4
|
||||
}
|
||||
}
|
||||
|
||||
fn fn02 (): int { }
|
||||
}"#,
|
||||
);
|
||||
assert_eq!(
|
||||
parser.parse_entity(),
|
||||
Some(Entity::Module(Module {
|
||||
name: "module01".into(),
|
||||
children: vec![
|
||||
ModuleChildren::Class(Class {
|
||||
name: "class01".into(),
|
||||
children: vec![ClassChildren::Fn(Fn {
|
||||
name: "fn01".into(),
|
||||
return_ty: None,
|
||||
params: vec![("param01".into(), Ty::Char), ("param02".into(), Ty::Float)],
|
||||
children: vec![Statement::Static(Let {
|
||||
name: "let01".into(),
|
||||
ty: Ty::Int,
|
||||
expr: Some(Expr::Literal(Literal::Int(4)))
|
||||
})]
|
||||
})]
|
||||
}),
|
||||
ModuleChildren::Fn(Fn {
|
||||
name: "fn02".into(),
|
||||
return_ty: Some(Ty::Int),
|
||||
params: vec![],
|
||||
children: vec![]
|
||||
})
|
||||
]
|
||||
}))
|
||||
);
|
||||
}
|
435
src/parser/expr.rs
Normal file
435
src/parser/expr.rs
Normal file
@@ -0,0 +1,435 @@
|
||||
use super::Parser;
|
||||
use crate::ast::{
|
||||
self,
|
||||
*,
|
||||
};
|
||||
use crate::lexer::{
|
||||
TokenDelimiter,
|
||||
TokenKeyword,
|
||||
TokenKind,
|
||||
TokenLiteral,
|
||||
TokenSymbol,
|
||||
};
|
||||
use std::rc::Rc;
|
||||
|
||||
impl<'a> Parser<'a> {
|
||||
/// exprIf ::= "if" expr block [ else (block | exprIf ) ]
|
||||
fn parse_expr_if(&mut self) -> Option<If> {
|
||||
// skip "if"
|
||||
self.next_token();
|
||||
|
||||
let cond = Box::new(self.parse_expr()?);
|
||||
let then = self.parse_expr_block()?;
|
||||
|
||||
if !self.skip_token(TokenKind::Keyword(TokenKeyword::Else)) {
|
||||
return Some(If {
|
||||
cond,
|
||||
then,
|
||||
or: None,
|
||||
});
|
||||
}
|
||||
|
||||
if self.peek_token().kind != TokenKind::Keyword(TokenKeyword::If) {
|
||||
return Some(If {
|
||||
cond,
|
||||
then,
|
||||
or: Some(Box::new(ElseType::Else(self.parse_expr_block()?))),
|
||||
});
|
||||
}
|
||||
|
||||
Some(If {
|
||||
cond,
|
||||
then,
|
||||
or: Some(Box::new(ElseType::If(self.parse_expr_if()?))),
|
||||
})
|
||||
}
|
||||
|
||||
/// exprBlock ::= "{" { statement } "}"
|
||||
fn parse_expr_block(&mut self) -> Option<Vec<Statement>> {
|
||||
let mut statements = vec![];
|
||||
|
||||
// skip {
|
||||
self.next_token();
|
||||
|
||||
loop {
|
||||
self.trim_newlines();
|
||||
if self.skip_token(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
break;
|
||||
}
|
||||
statements.push(self.parse_statement()?);
|
||||
if !self.check_newline_or_tok(TokenKind::Delimiter(TokenDelimiter::BraceClose)) {
|
||||
self.error_expected_peek("newline or }");
|
||||
return None;
|
||||
}
|
||||
}
|
||||
|
||||
Some(statements)
|
||||
}
|
||||
|
||||
/// exprLoop ::= "loop" exprBlock
|
||||
fn parse_expr_loop(&mut self) -> Option<Vec<Statement>> {
|
||||
self.next_token();
|
||||
if self.peek_token().kind != TokenKind::Delimiter(TokenDelimiter::BraceOpen) {
|
||||
self.error_expected_peek("{");
|
||||
return None;
|
||||
}
|
||||
|
||||
self.parse_expr_block()
|
||||
}
|
||||
|
||||
/// exprAtom ::= ( "(" expr ")" ) | ident | int | float | char | exprBlock | exprLoop | exprIf
|
||||
fn parse_expr_atom(&mut self) -> Option<Expr> {
|
||||
use ast::Literal::*;
|
||||
use TokenKind::*;
|
||||
|
||||
// TODO: check lvalue validity in the analysis phase
|
||||
Some(match self.peek_token().kind {
|
||||
Delimiter(TokenDelimiter::ParenOpen) => {
|
||||
self.next_token(); // skip (
|
||||
|
||||
let expr = self.parse_expr()?;
|
||||
|
||||
if !self.skip_token(TokenKind::Delimiter(TokenDelimiter::ParenClose)) {
|
||||
self.error_expected_peek(")");
|
||||
return None;
|
||||
}
|
||||
|
||||
expr
|
||||
}
|
||||
Identifier => {
|
||||
let token = self.next_token();
|
||||
Expr::Identifier(Rc::clone(&token.val))
|
||||
}
|
||||
Literal(TokenLiteral::Int) => Expr::Literal(Int(self.parse_int()?)),
|
||||
Literal(TokenLiteral::Float) => Expr::Literal(Float(self.parse_float()?)),
|
||||
Literal(TokenLiteral::Char) => Expr::Literal(Char(self.parse_char()?)),
|
||||
Delimiter(TokenDelimiter::BraceOpen) => Expr::Block(self.parse_expr_block()?),
|
||||
Keyword(TokenKeyword::Loop) => Expr::Loop(self.parse_expr_loop()?),
|
||||
Keyword(TokenKeyword::If) => Expr::If(self.parse_expr_if()?),
|
||||
_ => {
|
||||
self.error_expected_peek("expression");
|
||||
return None;
|
||||
}
|
||||
})
|
||||
}
|
||||
|
||||
/// exprUnary ::= [ unaryOp ] exprAtom
|
||||
/// unaryOp ::= "+" | "-" | "~"
|
||||
fn parse_expr_unary(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(symbol @ (Minus | Plus | Tilde)) => {
|
||||
self.next_token();
|
||||
Expr::Op(symbol, Box::new(self.parse_expr_atom()?), None)
|
||||
}
|
||||
_ => self.parse_expr_atom()?,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprArithmeticMul ::= exprUnary [ arithmeticMulOp exprArithmeticMul ]
|
||||
/// arithmeticMulOp ::= "*" | "/" | "%"
|
||||
fn parse_expr_arithmetic_mul(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
let lhs = self.parse_expr_unary()?;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(symbol @ (Star | Slash | Percent)) => {
|
||||
self.next_token();
|
||||
Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_arithmetic_mul()?)),
|
||||
)
|
||||
}
|
||||
_ => lhs,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprArithmeticAdd ::= exprArithmeticMul [ arithmeticAddOp exprArithmeticAdd ]
|
||||
/// arithmeticAddOp ::= "+" | "-"
|
||||
fn parse_expr_arithmetic_add(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
let lhs = self.parse_expr_arithmetic_mul()?;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(symbol @ (Plus | Minus)) => {
|
||||
self.next_token();
|
||||
Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_arithmetic_add()?)),
|
||||
)
|
||||
}
|
||||
_ => lhs,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprBitwiseShift ::= exprArithmeticAdd [ bitwiseShiftOp exprBitwiseShift ]
|
||||
/// bitwiseShiftOp ::= "<<" | ">>"
|
||||
fn parse_expr_bitwise_shift(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
let lhs = self.parse_expr_arithmetic_add()?;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(symbol @ (Shl | Shr)) => {
|
||||
self.next_token();
|
||||
Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_bitwise_shift()?)),
|
||||
)
|
||||
}
|
||||
_ => lhs,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprBitwiseAnd ::= exprBitwiseShift [ "&" exprBitwiseAnd ]
|
||||
fn parse_expr_bitwise_and(&mut self) -> Option<Expr> {
|
||||
let lhs = self.parse_expr_bitwise_shift()?;
|
||||
let symbol = TokenSymbol::And;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(symbol)) {
|
||||
return Some(lhs);
|
||||
}
|
||||
|
||||
Some(Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_bitwise_and()?)),
|
||||
))
|
||||
}
|
||||
|
||||
/// exprBitwiseXor ::= exprBitwiseAnd [ "^" exprBitwiseXor ]
|
||||
fn parse_expr_bitwise_xor(&mut self) -> Option<Expr> {
|
||||
let lhs = self.parse_expr_bitwise_and()?;
|
||||
let symbol = TokenSymbol::Caret;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(symbol)) {
|
||||
return Some(lhs);
|
||||
}
|
||||
|
||||
Some(Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_bitwise_xor()?)),
|
||||
))
|
||||
}
|
||||
|
||||
/// exprBiwiseOr ::= exprBitwiseXor [ "|" exprBitwiseOr ]
|
||||
fn parse_expr_bitwise_or(&mut self) -> Option<Expr> {
|
||||
let lhs = self.parse_expr_bitwise_xor()?;
|
||||
let symbol = TokenSymbol::Or;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(symbol)) {
|
||||
return Some(lhs);
|
||||
}
|
||||
|
||||
Some(Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_bitwise_or()?)),
|
||||
))
|
||||
}
|
||||
|
||||
/// exprAssign ::= exprBitwiseOr [ relationalOp exprRelational ]
|
||||
/// relationalOp ::= ">" | "<" | ">=" | "<=" | "==" | "!="
|
||||
fn parse_expr_relational(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
let lhs = self.parse_expr_bitwise_or()?;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(symbol @ (Gt | Lt | GtEq | LtEq | EqEq | Ne)) => {
|
||||
self.next_token();
|
||||
Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_relational()?)),
|
||||
)
|
||||
}
|
||||
_ => lhs,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprLogicalAnd ::= exprLogicalRelational [ "&&" exprLogicalAnd ]
|
||||
fn parse_expr_logical_and(&mut self) -> Option<Expr> {
|
||||
let lhs = self.parse_expr_relational()?;
|
||||
let symbol = TokenSymbol::AndAnd;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(symbol)) {
|
||||
return Some(lhs);
|
||||
}
|
||||
|
||||
Some(Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_logical_and()?)),
|
||||
))
|
||||
}
|
||||
|
||||
/// exprLogicalOr ::= exprLogicalAnd [ "||" exprLogicalOr ]
|
||||
fn parse_expr_logical_or(&mut self) -> Option<Expr> {
|
||||
let lhs = self.parse_expr_logical_and()?;
|
||||
let symbol = TokenSymbol::OrOr;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(symbol)) {
|
||||
return Some(lhs);
|
||||
}
|
||||
|
||||
Some(Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_logical_or()?)),
|
||||
))
|
||||
}
|
||||
|
||||
/// exprAssign ::= exprLogicalOr [ assignOp exprAssign ]
|
||||
/// assignOp ::= "=" | "+=" | "-=" | "*=" | "/=" | "%=" | "^=" | "<<=" | ">>=" | "&=" | "|="
|
||||
fn parse_expr_assign(&mut self) -> Option<Expr> {
|
||||
use TokenSymbol::*;
|
||||
let lhs = self.parse_expr_logical_or()?;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Symbol(
|
||||
symbol @ (Eq | PlusEq | MinusEq | StarEq | SlashEq | PercentEq | CaretEq | ShlEq
|
||||
| ShrEq | AndEq | OrEq),
|
||||
) => {
|
||||
self.next_token();
|
||||
Expr::Op(
|
||||
symbol,
|
||||
Box::new(lhs),
|
||||
Some(Box::new(self.parse_expr_assign()?)),
|
||||
)
|
||||
}
|
||||
_ => lhs,
|
||||
})
|
||||
}
|
||||
|
||||
/// exprControl ::= "continue" | "break" | "return" [ exprControl ] | exprAssign
|
||||
fn parse_expr_control(&mut self) -> Option<Expr> {
|
||||
use TokenKeyword::*;
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Keyword(Continue) => {
|
||||
self.next_token();
|
||||
Expr::Continue
|
||||
}
|
||||
TokenKind::Keyword(Break) => {
|
||||
self.next_token();
|
||||
Expr::Break
|
||||
}
|
||||
TokenKind::Keyword(Return) => {
|
||||
self.next_token();
|
||||
Expr::Return(self.parse_expr_control().map(Box::new))
|
||||
}
|
||||
_ => self.parse_expr_assign()?,
|
||||
})
|
||||
}
|
||||
|
||||
/// entrypoint for expression parsing using recursive descent parsing
|
||||
///
|
||||
/// <https://en.wikipedia.org/wiki/Recursive_descent_parser>
|
||||
/// expr ::= exprControl
|
||||
pub(super) fn parse_expr(&mut self) -> Option<Expr> {
|
||||
self.parse_expr_control()
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_parse_expr() {
|
||||
use Literal::*;
|
||||
use TokenSymbol::*;
|
||||
|
||||
macro_rules! b {
|
||||
($expr:expr) => {
|
||||
Box::new($expr)
|
||||
};
|
||||
}
|
||||
|
||||
let mut parser = Parser::new(
|
||||
r#"if if 1 { 1 } else { 0 } + 9 {
|
||||
a = 4
|
||||
} else if 1 {
|
||||
a = 5
|
||||
} else {
|
||||
}
|
||||
amul ^= (4 + 93 * (1 << 3) / 1.44) ^ bhatura
|
||||
stove = { 44 } + amul"#,
|
||||
);
|
||||
|
||||
assert_eq!(
|
||||
parser.parse_expr(),
|
||||
Some(Expr::If(If {
|
||||
cond: b!(Expr::Op(
|
||||
Plus,
|
||||
b!(Expr::If(If {
|
||||
cond: b!(Expr::Literal(Int(1))),
|
||||
then: vec![Statement::Expr(Expr::Literal(Int(1)))],
|
||||
or: Some(b!(ElseType::Else(vec![Statement::Expr(Expr::Literal(
|
||||
Int(0)
|
||||
))])))
|
||||
})),
|
||||
Some(b!(Expr::Literal(Int(9))))
|
||||
)),
|
||||
then: vec![Statement::Expr(Expr::Op(
|
||||
Eq,
|
||||
b!(Expr::Identifier("a".into())),
|
||||
Some(b!(Expr::Literal(Int(4))))
|
||||
))],
|
||||
or: Some(b!(ElseType::If(If {
|
||||
cond: b!(Expr::Literal(Int(1))),
|
||||
then: vec![Statement::Expr(Expr::Op(
|
||||
Eq,
|
||||
b!(Expr::Identifier("a".into())),
|
||||
Some(b!(Expr::Literal(Int(5))))
|
||||
))],
|
||||
or: Some(b!(ElseType::Else(vec![])))
|
||||
})))
|
||||
}))
|
||||
);
|
||||
|
||||
assert_eq!(parser.skip_token(TokenKind::Newline), true);
|
||||
|
||||
assert_eq!(
|
||||
parser.parse_expr(),
|
||||
Some(Expr::Op(
|
||||
CaretEq,
|
||||
b!(Expr::Identifier("amul".into())),
|
||||
Some(b!(Expr::Op(
|
||||
Caret,
|
||||
b!(Expr::Op(
|
||||
Plus,
|
||||
b!(Expr::Literal(Int(4))),
|
||||
Some(b!(Expr::Op(
|
||||
Star,
|
||||
b!(Expr::Literal(Int(93))),
|
||||
Some(b!(Expr::Op(
|
||||
Slash,
|
||||
b!(Expr::Op(
|
||||
Shl,
|
||||
b!(Expr::Literal(Int(1))),
|
||||
Some(b!(Expr::Literal(Int(3))))
|
||||
)),
|
||||
Some(b!(Expr::Literal(Float(1.44))))
|
||||
)))
|
||||
)))
|
||||
)),
|
||||
Some(b!(Expr::Identifier("bhatura".into())))
|
||||
)))
|
||||
))
|
||||
);
|
||||
|
||||
assert_eq!(parser.skip_token(TokenKind::Newline), true);
|
||||
|
||||
assert_eq!(
|
||||
parser.parse_expr(),
|
||||
Some(Expr::Op(
|
||||
Eq,
|
||||
b!(Expr::Identifier("stove".into())),
|
||||
Some(b!(Expr::Op(
|
||||
Plus,
|
||||
b!(Expr::Block(vec![Statement::Expr(Expr::Literal(Int(44)))])),
|
||||
Some(b!(Expr::Identifier("amul".into())))
|
||||
)))
|
||||
))
|
||||
);
|
||||
}
|
121
src/parser/literal.rs
Normal file
121
src/parser/literal.rs
Normal file
@@ -0,0 +1,121 @@
|
||||
use super::Parser;
|
||||
use crate::lexer::{
|
||||
TokenKind,
|
||||
TokenLiteral,
|
||||
TokenSymbol,
|
||||
};
|
||||
|
||||
impl<'a> Parser<'a> {
|
||||
/// int ::= digit { digit }
|
||||
pub(super) fn parse_int(&mut self) -> Option<i32> {
|
||||
let val = self.next_token().val;
|
||||
let mut integer: i32 = 0;
|
||||
let error = || {
|
||||
self.error(&format!(
|
||||
"integer values must be in range [{}, {}]",
|
||||
i32::MIN,
|
||||
i32::MAX
|
||||
))
|
||||
};
|
||||
|
||||
for c in val.chars() {
|
||||
// c is always ['0'..='9']
|
||||
let d = c.to_digit(10)?;
|
||||
|
||||
match integer.checked_mul(10) {
|
||||
Some(m) => integer = m,
|
||||
None => {
|
||||
error();
|
||||
return None;
|
||||
}
|
||||
}
|
||||
|
||||
match integer.checked_add(d as i32) {
|
||||
Some(a) => integer = a,
|
||||
None => {
|
||||
error();
|
||||
return None;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
Some(integer)
|
||||
}
|
||||
|
||||
// didnt use parse() because i wanted to do this myself for some reason
|
||||
/// f32 can be NaN and inf as well
|
||||
/// float ::= int [ "." { digit } ] [ "e" { digit } ]
|
||||
pub(super) fn parse_float(&mut self) -> Option<f32> {
|
||||
let token = self.next_token();
|
||||
let mut chars = token.val.chars();
|
||||
let mut float: f32 = 0.0;
|
||||
let mut fraction: f32 = 0.0;
|
||||
let mut prec: i32 = 0;
|
||||
let mut exp: i32 = 0;
|
||||
let mut decimal: bool = false;
|
||||
|
||||
// lexer takes care of multiple decimals and non digit characters
|
||||
for c in chars.by_ref() {
|
||||
match c {
|
||||
'.' => decimal = true,
|
||||
'e' | 'E' => {
|
||||
// lexer takes care that decimal doesnt come after e
|
||||
let s;
|
||||
match self.peek_token().kind {
|
||||
TokenKind::Symbol(TokenSymbol::Minus) => {
|
||||
s = -1;
|
||||
self.next_token();
|
||||
}
|
||||
TokenKind::Symbol(TokenSymbol::Plus) => {
|
||||
s = 1;
|
||||
self.next_token();
|
||||
}
|
||||
_ => s = 1,
|
||||
}
|
||||
|
||||
if self.peek_token().kind != TokenKind::Literal(TokenLiteral::Int) {
|
||||
break;
|
||||
}
|
||||
|
||||
exp = self.parse_int()? * s;
|
||||
break;
|
||||
}
|
||||
_ => {
|
||||
// c is always ['0'..='9']
|
||||
let d = c.to_digit(10)? as f32;
|
||||
if decimal {
|
||||
fraction *= 10.0;
|
||||
fraction += d;
|
||||
prec += 1;
|
||||
} else {
|
||||
float *= 10.0;
|
||||
float += d;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
fraction /= 10f32.powi(prec);
|
||||
float += fraction;
|
||||
float *= 10f32.powi(exp);
|
||||
|
||||
Some(float)
|
||||
}
|
||||
|
||||
/// char ::= "'" letter "'"
|
||||
pub(super) fn parse_char(&mut self) -> Option<char> {
|
||||
// the lexer ensures that the 0th and 2nd characters are both '
|
||||
self.next_token().val.chars().nth(1)
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_parse_literals() {
|
||||
let mut parser = Parser::new("4524 3123.15e4 9e2 9083482.429455 'c' 3331.13.1");
|
||||
assert_eq!(parser.parse_int(), Some(4524));
|
||||
assert_eq!(parser.parse_float(), Some(3123.15e4));
|
||||
assert_eq!(parser.parse_float(), Some(9e2));
|
||||
assert_eq!(parser.parse_float(), Some(9083482.429455));
|
||||
assert_eq!(parser.parse_char(), Some('c'));
|
||||
assert_eq!(parser.next_token().kind, TokenKind::Invalid);
|
||||
}
|
156
src/parser/mod.rs
Normal file
156
src/parser/mod.rs
Normal file
@@ -0,0 +1,156 @@
|
||||
//! A naive parser just to get started
|
||||
//!
|
||||
//! Can only parse module, class and function declaration now along with let statements
|
||||
|
||||
mod entity;
|
||||
mod expr;
|
||||
mod literal;
|
||||
mod statement;
|
||||
|
||||
use crate::ast::{
|
||||
Parent,
|
||||
Ty,
|
||||
};
|
||||
use crate::lexer::{
|
||||
Lexer,
|
||||
Token,
|
||||
TokenKeyword,
|
||||
TokenKind,
|
||||
TokenSymbol,
|
||||
};
|
||||
use std::rc::Rc;
|
||||
|
||||
pub struct Parser<'a> {
|
||||
pub lexer: Lexer<'a>,
|
||||
}
|
||||
|
||||
impl<'a> Parser<'a> {
|
||||
/// Creates a new [`Parser`] instance.
|
||||
pub fn new(contents: &'a str) -> Parser<'a> {
|
||||
Parser {
|
||||
lexer: Lexer::new(contents),
|
||||
}
|
||||
}
|
||||
|
||||
#[inline]
|
||||
fn error(&self, message: &str) {
|
||||
eprintln!(
|
||||
"Parser: {}, at \"{}:{}\"",
|
||||
message, self.lexer.line, self.lexer.col
|
||||
);
|
||||
}
|
||||
|
||||
#[inline]
|
||||
fn error_expected(&self, expected: &str, found: &str) {
|
||||
self.error(&format!("expected {}, found {}", expected, found));
|
||||
}
|
||||
|
||||
#[inline]
|
||||
fn error_expected_peek(&mut self, expected: &str) {
|
||||
let found = &Rc::clone(&self.peek_token().val);
|
||||
self.error_expected(expected, found);
|
||||
}
|
||||
|
||||
#[inline]
|
||||
fn next_token(&mut self) -> Token {
|
||||
let t = self.lexer.next_token();
|
||||
println!("{:?}", t);
|
||||
t
|
||||
}
|
||||
|
||||
#[inline]
|
||||
fn peek_token(&mut self) -> &Token {
|
||||
return self.lexer.peek_token();
|
||||
}
|
||||
|
||||
/// newline ::= "}\n"
|
||||
fn trim_newlines(&mut self) {
|
||||
while self.peek_token().kind == TokenKind::Newline {
|
||||
self.next_token();
|
||||
}
|
||||
}
|
||||
|
||||
fn skip_token(&mut self, kind: TokenKind) -> bool {
|
||||
if self.peek_token().kind == kind {
|
||||
self.next_token();
|
||||
return true;
|
||||
}
|
||||
|
||||
false
|
||||
}
|
||||
|
||||
fn check_newline_or_tok(&mut self, token: TokenKind) -> bool {
|
||||
match self.peek_token().kind {
|
||||
TokenKind::Newline => true,
|
||||
d if d == token => true,
|
||||
_ => false,
|
||||
}
|
||||
}
|
||||
|
||||
/// ty ::= "int" | "float" | "char"
|
||||
fn parse_ty(&mut self) -> Option<Ty> {
|
||||
let ty: Ty;
|
||||
|
||||
if let TokenKind::Keyword(keyword) = &self.peek_token().kind {
|
||||
ty = match keyword {
|
||||
TokenKeyword::Int => Ty::Int,
|
||||
TokenKeyword::Char => Ty::Char,
|
||||
TokenKeyword::Float => Ty::Float,
|
||||
_ => {
|
||||
self.error_expected_peek("ty");
|
||||
return None;
|
||||
}
|
||||
};
|
||||
} else {
|
||||
self.error_expected_peek("ty");
|
||||
return None;
|
||||
}
|
||||
self.next_token();
|
||||
Some(ty)
|
||||
}
|
||||
|
||||
/// ident ::= ( letter | "_" ) { letter | digit | "_" }
|
||||
fn parse_ident(&mut self) -> Option<Rc<str>> {
|
||||
if self.peek_token().kind != TokenKind::Identifier {
|
||||
self.error_expected_peek("identifier");
|
||||
return None;
|
||||
}
|
||||
|
||||
Some(Rc::clone(&self.next_token().val))
|
||||
}
|
||||
|
||||
/// identWithTy ::= letter ":" ty
|
||||
fn parse_ident_with_ty(&mut self) -> Option<(Rc<str>, Ty)> {
|
||||
let ident = self.parse_ident()?;
|
||||
|
||||
if !self.skip_token(TokenKind::Symbol(TokenSymbol::Colon)) {
|
||||
self.error_expected_peek(":");
|
||||
return None;
|
||||
}
|
||||
|
||||
Some((ident, self.parse_ty()?))
|
||||
}
|
||||
|
||||
/// Returns an [`Entity`] vector after parsing
|
||||
///
|
||||
/// parent ::= { entity }
|
||||
/// [`Entity`]: crate::ast::Entity
|
||||
pub fn parse(&mut self) -> Option<Parent> {
|
||||
let mut parent = vec![];
|
||||
|
||||
loop {
|
||||
match self.peek_token().kind {
|
||||
TokenKind::Newline => self.trim_newlines(),
|
||||
TokenKind::Eof => break,
|
||||
_ => {
|
||||
parent.push(self.parse_entity()?);
|
||||
if !self.check_newline_or_tok(TokenKind::Eof) {
|
||||
self.error_expected_peek("newline or end of file");
|
||||
return None;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
Some(parent)
|
||||
}
|
||||
}
|
91
src/parser/statement.rs
Normal file
91
src/parser/statement.rs
Normal file
@@ -0,0 +1,91 @@
|
||||
use super::Parser;
|
||||
use crate::ast::*;
|
||||
use crate::lexer::{
|
||||
TokenKeyword,
|
||||
TokenKind,
|
||||
TokenSymbol,
|
||||
};
|
||||
|
||||
impl<'a> Parser<'a> {
|
||||
/// statement ::= static | let | expr
|
||||
pub(super) fn parse_statement(&mut self) -> Option<Statement> {
|
||||
use TokenKeyword::*;
|
||||
println!("STMT");
|
||||
|
||||
Some(match self.peek_token().kind {
|
||||
TokenKind::Keyword(Static) => Statement::Static(self.parse_static()?),
|
||||
TokenKind::Keyword(Let) => Statement::Let(self.parse_let()?),
|
||||
_ => Statement::Expr(self.parse_expr()?),
|
||||
})
|
||||
}
|
||||
|
||||
/// static ::="static" let
|
||||
pub(super) fn parse_static(&mut self) -> Option<Let> {
|
||||
self.next_token();
|
||||
|
||||
if self.peek_token().kind != TokenKind::Keyword(TokenKeyword::Let) {
|
||||
self.error_expected_peek("let");
|
||||
return None;
|
||||
}
|
||||
|
||||
self.parse_let()
|
||||
}
|
||||
|
||||
/// let ::= "let" identWithTy "=" expr
|
||||
pub(super) fn parse_let(&mut self) -> Option<Let> {
|
||||
self.next_token();
|
||||
|
||||
let (name, ty) = self.parse_ident_with_ty()?;
|
||||
|
||||
let expr = if self.skip_token(TokenKind::Symbol(TokenSymbol::Eq)) {
|
||||
self.parse_expr()
|
||||
} else if self.peek_token().kind == TokenKind::Newline {
|
||||
None
|
||||
} else {
|
||||
self.error_expected_peek("= or newline");
|
||||
return None;
|
||||
};
|
||||
|
||||
Some(Let { name, ty, expr })
|
||||
}
|
||||
}
|
||||
|
||||
#[test]
|
||||
fn test_parse_let() {
|
||||
use Literal::*;
|
||||
|
||||
let mut parser = Parser::new(
|
||||
r#"static let test01: int = 4
|
||||
let test02: char = '6'
|
||||
static let test03: float
|
||||
let test04 = 9"#,
|
||||
);
|
||||
assert_eq!(
|
||||
parser.parse_static(),
|
||||
Some(Let {
|
||||
name: "test01".into(),
|
||||
ty: Ty::Int,
|
||||
expr: Some(Expr::Literal(Int(4)))
|
||||
})
|
||||
);
|
||||
assert_eq!(parser.skip_token(TokenKind::Newline), true);
|
||||
assert_eq!(
|
||||
parser.parse_let(),
|
||||
Some(Let {
|
||||
name: "test02".into(),
|
||||
ty: Ty::Char,
|
||||
expr: Some(Expr::Literal(Char('6')))
|
||||
})
|
||||
);
|
||||
assert_eq!(parser.skip_token(TokenKind::Newline), true);
|
||||
assert_eq!(
|
||||
parser.parse_static(),
|
||||
Some(Let {
|
||||
name: "test03".into(),
|
||||
ty: Ty::Float,
|
||||
expr: None
|
||||
})
|
||||
);
|
||||
assert_eq!(parser.skip_token(TokenKind::Newline), true);
|
||||
assert_eq!(parser.parse_let(), None);
|
||||
}
|
Reference in New Issue
Block a user