Compare commits

...

9 Commits

Author SHA1 Message Date
ondra05 2bf8445a5f
Multiple things 2023-10-04 19:02:02 +02:00
ondra05 e9ebc89a2d
New parser 2023-10-04 18:59:44 +02:00
ondra05 814059fda6
Added some syntax 2023-10-04 02:38:27 +02:00
ondra05 114adce336
Macro 2023-10-03 23:28:57 +02:00
ondra05 f4895be6b1
AST goes brrrr 2023-10-03 02:20:25 +02:00
ondra05 e87c3ab904
Add KW 2023-10-03 01:44:15 +02:00
ondra05 376cd60a4f
Lexer texts 2023-10-03 01:38:44 +02:00
ondra05 7da236db3f
Added lexer 2023-10-03 01:21:47 +02:00
ondra05 9345932306
Removed syntax 2023-10-03 00:30:31 +02:00
7 changed files with 284 additions and 319 deletions

28
Cargo.lock generated
View File

@ -80,6 +80,26 @@ version = "0.2.148"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "9cdc71e17332e86d2e1d38c1f99edcb6288ee11b815fb1a4b049eaa2114d369b"
[[package]]
name = "literify"
version = "0.2.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "fd552332051e9b3db140d34a371dcc0ed378b72a9227b5273070af58ea34abf4"
dependencies = [
"litrs",
"proc-macro2",
"quote",
]
[[package]]
name = "litrs"
version = "0.4.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "4f17c3668f3cc1132437cdadc93dab05e52d592f06948d3f64828430c36e4a70"
dependencies = [
"proc-macro2",
]
[[package]]
name = "logos"
version = "0.13.0"
@ -118,6 +138,12 @@ version = "1.18.0"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "dd8b5dd2ae5ed71462c540258bedcb51965123ad7e7ccf4b9a8cafaa4a63576d"
[[package]]
name = "paste"
version = "1.0.14"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "de3145af08024dea9fa9914f381a17b8fc6034dfb00f3a84013f7ff43f29ed4c"
[[package]]
name = "proc-macro2"
version = "1.0.67"
@ -158,7 +184,9 @@ dependencies = [
"bumpalo",
"chumsky",
"lasso",
"literify",
"logos",
"paste",
]
[[package]]

View File

@ -4,7 +4,9 @@ version = "0.1.0"
edition = "2021"
[dependencies]
bumpalo = { version = "3", features = ["collections"] }
chumsky = "1.0.0-alpha"
lasso = "0.7"
logos = "0.13"
bumpalo = { version = "3", features = ["collections"] }
chumsky = "1.0.0-alpha"
lasso = "0.7"
literify = "0.2"
logos = "0.13"
paste = "1.0"

3
rustfmt.toml Normal file
View File

@ -0,0 +1,3 @@
enum_discrim_align_threshold = 16
struct_field_align_threshold = 16
imports_granularity = "one"

View File

@ -1,21 +1,22 @@
// Rhea
use bumpalo::Bump;
use logos::Logos;
use std::io::{stdin, Read};
use utils::default;
use {logos::Logos, syntax::token::Token};
mod syntax;
mod utils;
use {
bumpalo::Bump,
std::io::{stdin, Read},
utils::default,
};
fn main() -> Result<(), Box<dyn std::error::Error>> {
let mut buf = default();
stdin().read_to_string(&mut buf)?;
let lexer = syntax::token::Token::lexer_with_extras(&buf, default());
let arena = Bump::new();
syntax::parser::parse_lexer(lexer, &arena);
syntax::parser::parse_lexer(Token::lexer(&buf), &arena);
Ok(())
}

View File

@ -1,4 +1,4 @@
use {super::token::IntLit, chumsky::span::SimpleSpan, lasso::Spur};
use {chumsky::span::SimpleSpan, lasso::Spur};
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub struct Spanned<T> {
@ -8,9 +8,16 @@ pub struct Spanned<T> {
impl<T> Spanned<T> {
#[inline]
pub fn new(item: T, span: SimpleSpan) -> Self {
pub const fn new(item: T, span: SimpleSpan) -> Self {
Self { item, span }
}
pub fn map<U>(self, mut f: impl FnMut(T) -> U) -> Spanned<U> {
Spanned {
item: f(self.item),
span: self.span,
}
}
}
pub type SpanExpr<'a> = Spanned<Expr<'a>>;
@ -24,60 +31,82 @@ pub enum Expr<'a> {
Path(ExprList<'a>),
Literal(Literal),
Call(ExprRef<'a>, ExprList<'a>),
Binary(Spanned<BinaryOperator>, ExprRef<'a>, ExprRef<'a>),
Unary(Spanned<UnaryOperator>, ExprRef<'a>),
BindLocal(Spanned<Pattern>, ExprRef<'a>, Option<ExprList<'a>>),
BindIn(
Spanned<Pattern>,
ExprRef<'a>,
ExprList<'a>,
Option<ExprList<'a>>,
),
Binary(Spanned<BinaryOp>, ExprRef<'a>, ExprRef<'a>),
Unary(Spanned<UnaryOp>, ExprRef<'a>),
Set(ExprRef<'a>, ExprRef<'a>),
Match(ExprRef<'a>, &'a [(Spanned<Pattern>, SpanExpr<'a>)]),
Func(&'a [(Spanned<Pattern>, Spanned<Type>)], Spanned<Type>, ExprRef<'a>),
Loop(ExprList<'a>),
Block(ExprList<'a>),
Unit,
Definition(Definition<'a>),
Switch {
on: ExprRef<'a>,
branches: &'a [(Spanned<Switcher>, ExprRef<'a>)],
else_: ExprRef<'a>,
},
CondSwitch {
branches: &'a [(ExprRef<'a>, ExprRef<'a>)],
else_: ExprRef<'a>,
},
Break(Option<ExprRef<'a>>),
Return(Option<ExprRef<'a>>),
Uninit,
Continue,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum Definition<'a> {
Binding {
kind: DefKind,
ident: Spanned<Ident>,
ty: Option<Spanned<Type>>,
init: ExprRef<'a>,
},
Func {
name: Spanned<Ident>,
params: &'a [(Spanned<Ident>, Spanned<Type>)],
ret: Spanned<Type>,
body: ExprList<'a>,
},
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum Type {
Ident(Ident),
Unit,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum Pattern {
Ident(Ident),
pub enum Switcher {
Literal(Literal),
None,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum DefKind {
Const,
Var,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum Literal {
String(Spur),
Integer(IntLit),
Int(u64),
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum BinaryOperator {
pub enum BinaryOp {
Plus,
Minus,
Star,
Slash,
And,
VLine,
Equ,
Neq,
Lt,
Gt,
Equ,
Nequ,
LtEqu,
GtEqu,
LtEq,
GtEq,
}
#[derive(Clone, Copy, Debug, PartialEq, Eq)]
pub enum UnaryOperator {
Tilde,
pub enum UnaryOp {
Minus,
Not,
Star,
}

View File

@ -1,9 +1,9 @@
use super::ast::Type;
use super::ast::{DefKind, Expr, ExprList, Type};
use {
super::{
ast::{BinaryOperator, Expr, Literal, Pattern, SpanExpr, Spanned, UnaryOperator},
token::Token,
ast::{Definition, Ident, SpanExpr, Spanned},
token::{Token, T},
},
crate::utils::Pipe,
bumpalo::Bump,
@ -23,195 +23,70 @@ macro_rules! equivmap {
};
}
fn expr<'a, I>() -> impl Parser<'a, I, SpanExpr<'a>, Extra<'a>> + Clone
fn ident<'a, I>() -> impl Parser<'a, I, Spanned<Ident>, Extra<'a>> + Clone + Copy
where
I: Input<'a, Token = Token, Span = SimpleSpan> + ValueInput<'a>,
{
recursive(|expr| {
let ident = select!(Token::Ident(id) => id);
select!(Token::Ident(id) => id).map_with_span(Spanned::new)
}
let literal = select! {
Token::Int(a) => Literal::Integer(a),
Token::String(a) => Literal::String(a)
};
fn ty<'a, I>() -> impl Parser<'a, I, Spanned<Type>, Extra<'a>> + Clone + Copy
where
I: Input<'a, Token = Token, Span = SimpleSpan> + ValueInput<'a>,
{
ident().map(|i| i.map(Type::Ident))
}
let pattern = select! {
Token::Ident(id) => Pattern::Ident(id),
Token::Underscore => Pattern::None,
}
.or(literal.map(Pattern::Literal))
.map_with_span(Spanned::new);
fn definition<'a, I>() -> impl Parser<'a, I, Spanned<Definition<'a>>, Extra<'a>> + Clone
where
I: Input<'a, Token = Token, Span = SimpleSpan> + ValueInput<'a>,
{
let ident = ident();
let ty = ty();
let type_ = just([Token::LeftParen, Token::RightParen])
.to(Type::Unit)
.or(ident.map(Type::Ident))
.map_with_span(Spanned::new);
let block = expr
.clone()
.separated_by(just(Token::Semicolon))
.allow_trailing()
.pipe(arena_collect)
.delimited_by(just(Token::LeftCurly), just(Token::RightCurly));
let func = just(Token::Func)
.ignore_then(
pattern
.then_ignore(just(Token::Colon))
.then(type_)
.separated_by(just(Token::Comma))
.allow_trailing()
.pipe(arena_collect)
.delimited_by(just(Token::LeftParen), just(Token::RightParen)),
)
.then_ignore(just(Token::Colon))
.then(type_)
.then(
just(Token::Equ)
.ignore_then(expr.clone())
.or(block.clone().map(Expr::Block).map_with_span(Spanned::new)),
)
.map_with_state(|((params, ret), expr), _, state| {
Expr::Func(params, ret, state.arena.alloc(expr))
});
let atom = literal
.map(Expr::Literal)
.or(just([Token::LeftParen, Token::RightParen]).to(Expr::Unit))
.or(ident.map(Expr::Ident))
.or(func)
.map_with_span(Spanned::new)
.or(expr
.clone()
.delimited_by(just(Token::LeftParen), just(Token::RightParen)));
// <expr>(expr1, expr2, …)
let call = atom.clone().foldl_with_state(
expr.clone()
.separated_by(just(Token::Comma))
let func = just(T!["func"])
.ignore_then(ident)
.then(
ident
.then_ignore(just(T![":"]))
.then(ty)
.separated_by(just(T![","]))
.allow_trailing()
.pipe(arena_collect)
.delimited_by(just(Token::LeftParen), just(Token::RightParen))
.map_with_span(Spanned::new)
.repeated(),
|expr, paramlist, state: &mut State| {
Spanned::new(
Expr::Call(state.arena.alloc(expr), paramlist.item),
merge_spans(expr.span, paramlist.span),
)
.delimited_by(just(T!["("]), just(T![")"])),
)
.then_ignore(just(T![""]))
.then(ty)
.then(just([T!["{"], T!["}"]]))
.map_with_state(
|(((name, params), ret), _body), _, state| Definition::Func {
name,
params,
ret,
body: state.arena.alloc_slice_copy(&[]),
},
);
let path = call
.clone()
.map_with_state(|item, _, state| bumpalo::vec![in state.arena; item])
.foldl(
just(Token::Dot).ignore_then(call).repeated(),
|mut v, expr| {
v.push(expr);
v
},
)
.map(|v| Expr::Path(v.into_bump_slice()))
.map_with_span(Spanned::new);
/* let unary = equivmap!(Token, UnaryOperator, [Minus, Tilde])
.map_with_span(Spanned::new)
.repeated()
.foldr_with_state(call, |op, expr, state| {
Spanned::new(
Expr::Unary(op, state.arena.alloc(expr)),
merge_spans(op.span, expr.span),
)
});
*/
let unary = path.foldl_with_state(
just([Token::Dot, Token::Star])
.to(UnaryOperator::Star)
.or(just(Token::Tilde).to(UnaryOperator::Tilde))
.map_with_span(Spanned::new)
.repeated(),
|expr, op, state| {
Spanned::new(
Expr::Unary(op, state.arena.alloc(expr)),
merge_spans(expr.span, op.span),
)
let binding = equivmap!(Token, DefKind, [Const, Var])
.then(ident)
.then(just(T![":"]).ignore_then(ty).or_not())
.then(
just(T!["="]).ignore_then(
just(T!["uninit"])
.to(Expr::Uninit)
.map_with_span(Spanned::new),
),
)
.map_with_state(
|(((kind, ident), ty), init), _, state| Definition::Binding {
kind,
ident,
ty,
init: state.arena.alloc(init),
},
);
// <exprL> OP <exprR>
let binary = unary.clone().foldl_with_state(
equivmap!(
Token,
BinaryOperator,
[Plus, Minus, Star, Slash, And, VLine, Lt, Gt, Equ, Nequ, LtEqu, GtEqu],
)
.map_with_span(Spanned::new)
.then(unary)
.repeated(),
|l, (op, r), state: &mut State| {
Spanned::new(
Expr::Binary(op, state.arena.alloc(l), state.arena.alloc(r)),
merge_spans(l.span, r.span),
)
},
);
let bind = {
let start = pattern.then_ignore(just(Token::Colon)).then(expr.clone()); // <pat> := <expr>
let else_ = just(Token::Else).ignore_then(block.clone()).or_not(); // else {…}
// <pat> := <expr> [else {…}]
let local = start.clone().then(else_.clone()).map_with_state(
|((pat, expr), else_), _, state| {
Expr::BindLocal(pat, &*state.arena.alloc(expr), else_)
},
);
// <pat> := <expr> {…} else {…}
let in_ = start.then(block.clone()).then(else_).map_with_state(
|(((pat, expr), block), else_), _, state| {
Expr::BindIn(pat, &*state.arena.alloc(expr), block, else_)
},
);
in_.or(local)
};
// <atom> ← <expr>
let set = atom
.clone()
.then_ignore(just(Token::LArrow))
.then(expr.clone())
.map_with_state(|(place, expr), _, state| {
Expr::Set(state.arena.alloc(place), state.arena.alloc(expr))
});
// <expr>.match { <pat> → <expr>, … }
let match_ = atom
.clone()
.then_ignore(just([Token::Dot, Token::Match]))
.then(
pattern
.then_ignore(just(Token::RArrow))
.then(expr)
.separated_by(just(Token::Comma))
.allow_trailing()
.pipe(arena_collect)
.delimited_by(just(Token::LeftCurly), just(Token::RightCurly)),
)
.map_with_state(|(expr, branches), _, state| {
Expr::Match(state.arena.alloc(expr), branches)
});
bind.or(set)
.or(match_)
.or(block.map(Expr::Block))
.map_with_span(Spanned::new)
.or(binary)
.or(atom)
})
func.or(binding).map_with_span(Spanned::new)
}
pub struct State<'a> {
@ -227,8 +102,8 @@ pub fn parse_input<'a>(
) -> ParseResult {
println!(
"{:?}",
expr()
.separated_by(just(Token::Semicolon))
definition()
.separated_by(just(T![";"]))
.allow_trailing()
.pipe(arena_collect)
.parse_with_state(input, &mut State { arena })

View File

@ -1,112 +1,139 @@
use lasso::Spur;
use logos::Lexer;
use {lasso::Rodeo, logos::Logos};
use lasso::{Rodeo, Spur};
use logos::{Lexer, Logos};
#[derive(Default)]
pub struct Lextras {
pub struct Extras {
pub interner: Rodeo,
}
#[derive(Copy, Clone, Debug, PartialEq, Eq)]
pub enum IntLit {
Signed(i64),
Unsigned(u64),
macro_rules! token_def {
(
unit { $($u_name:ident : $($u_tok:literal),* $(,)?;)* }
keyword { $($kw:tt),* $(,)* }
else { $($e_tt:tt)* }
) => {
literify::literify!(paste::paste! {
#[derive(Clone, Copy, Debug, PartialEq, Eq, Logos)]
#[logos(extras = Extras)]
#[logos(skip r"[ \t\n\f]+")]
#[logos(skip r"\\.*")]
pub enum Token {
$(
$(#[token($u_tok)])*
$u_name,
)*
$(
#[token(~($kw))]
[<$kw:camel>],
)*
$($e_tt)*
}
macro_rules! T {
$($(
($u_tok) => { $crate::syntax::token::Token::$u_name };
)*)*
$((~($kw)) => { $crate::syntax::token::Token::[<$kw:camel>] };)*
}
});
};
}
#[derive(Logos, Copy, Clone, Debug, PartialEq, Eq)]
#[logos(extras = Lextras)]
#[logos(skip r"[ \t\n\f]+")]
#[logos(skip r"-- .*")]
#[rustfmt::skip]
pub enum Token {
#[token("(")] LeftParen,
#[token(")")] RightParen,
#[token("{")] LeftCurly,
#[token("}")] RightCurly,
#[token(".")] Dot,
#[token(",")] Comma,
#[token(":")] Colon,
#[token(";")] Semicolon,
#[token("_")] Underscore,
token_def!(
unit {
LeftParen : "(";
RightParen: ")";
LeftCurly : "{";
RightCurly: "}";
Dot : ".";
Comma : ",";
Colon : ":";
Semicolon : ";";
#[token("")] //____
#[token("<-")] LArrow,
#[token("")] //____
#[token("->")] RArrow,
LArrow: "", "<-";
RArrow: "", "->";
#[token(":>")] Pipe,
Plus : "+";
Minus : "-";
Star : "*";
Slash : "/";
Precent: "%";
#[token("+")] Plus,
#[token("-")] Minus,
#[token("*")] Star,
#[token("/")] Slash,
#[token("&")] And,
#[token("|")] VLine,
#[token("~")] Tilde,
#[token("<")] Lt,
#[token(">")] Gt,
#[token("=")] Equ,
#[token("") ] //__
#[token("/=")] Nequ,
#[token("") ] //___
#[token("<=")] LtEqu,
#[token("") ] //___,
#[token(">=")] GtEqu,
#[token("match")] Match,
#[token("else")] Else,
#[token("loop")] Loop,
#[token("const")] Const,
#[token("var")] Var,
#[token("func")] Func,
// Modules aren't real here ondra just variables with imported functions
#[token("module")] Module,
#[regex(
r"\p{XID_Start}\p{XID_Continue}*",
|l| l.extras.interner.get_or_intern(l.slice())
)] Ident(Spur),
#[token("»", better_string)]
#[regex(
"\"[^\"]*\"",
|l| {
let slice = l.slice();
l.extras.interner.get_or_intern(&slice[1..slice.len() - 1])
}
)] String(Spur),
#[regex(
"-?[0-9]+",
|l| {
Some(if let Some(slice) = l.slice().strip_prefix('-') {
IntLit::Signed(slice.parse::<i64>().ok()?)
} else {
IntLit::Unsigned(l.slice().parse::<u64>().ok()?)
})
}
)] Int(IntLit),
Invalid,
}
// For Evy, with love.
fn better_string(lexer: &mut Lexer<Token>) -> Option<Spur> {
let mut count = 1;
for (ix, chr) in lexer.remainder().char_indices() {
match chr {
'«' => count -= 1,
'»' => count += 1,
_ => (),
}
if count == 0 {
let slice = &lexer.remainder()[..ix];
lexer.bump(ix + '«'.len_utf8());
return Some(lexer.extras.interner.get_or_intern(slice));
}
Equ : "=";
Neq : "", "/=";
Lt : "<";
Gt : ">";
LtEq : "", "<=";
GtEq : "", ">=";
}
keyword { func, var, const, include, switch, loop,
return, break, continue, uninit, asm }
else {
#[regex(
r"\p{XID_Start}\p{XID_Continue}*",
|l| intern(l, l.slice()),
)] Ident(Spur),
#[regex(
"\"[^\"]*\"",
|l| {
let s = l.slice();
intern(l, &s[1..s.len() - 1])
},
)] String(Spur),
#[regex(
"[0-9]+",
|l| l.slice().parse::<u64>().ok()
)] Int(u64),
Invalid,
}
);
pub(crate) use T;
fn intern(lexer: &mut Lexer<'_, Token>, s: &str) -> Spur {
lexer.extras.interner.get_or_intern(s)
}
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn ident() {
let mut lexer = Token::lexer("いえぶる able");
assert_eq!(
lexer.next(),
Some(Ok(Token::Ident(intern(&mut lexer, "いえぶる")))),
);
assert_eq!(
lexer.next(),
Some(Ok(Token::Ident(intern(&mut lexer, "able")))),
);
}
#[test]
fn string() {
let mut lexer = Token::lexer("\"sussy\" \"baka");
assert_eq!(
lexer.next(),
Some(Ok(Token::String(intern(&mut lexer, "sussy")))),
);
assert_eq!(lexer.next(), Some(Err(())),);
}
#[test]
fn symbol_alt() {
let mut lexer = Token::lexer("-> →");
assert_eq!(lexer.next(), lexer.next());
}
None
}