Multiple things

Cargo.lock

@@ -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]]

Cargo.toml

@@ -7,4 +7,6 @@ edition = "2021"
 bumpalo  = { version = "3", features = ["collections"] }
 chumsky  = "1.0.0-alpha"
 lasso    = "0.7"
+literify = "0.2"
 logos    = "0.13"
+paste    = "1.0"

rustfmt.toml

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

src/main.rs

@@ -1,21 +1,22 @@
 // Rhea
 
-use bumpalo::Bump;
+use {logos::Logos, syntax::token::Token};
-use logos::Logos;
-use std::io::{stdin, Read};
-use utils::default;
 
 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(())
 }

src/syntax/ast.rs

@@ -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>),
+    Binary(Spanned<BinaryOp>, ExprRef<'a>, ExprRef<'a>),
-    Unary(Spanned<UnaryOperator>, ExprRef<'a>),
+    Unary(Spanned<UnaryOp>, ExprRef<'a>),
-    BindLocal(Spanned<Pattern>, ExprRef<'a>, Option<ExprList<'a>>),
-    BindIn(
-        Spanned<Pattern>,
-        ExprRef<'a>,
-        ExprList<'a>,
-        Option<ExprList<'a>>,
-    ),
     Set(ExprRef<'a>, ExprRef<'a>),
-    Match(ExprRef<'a>, &'a [(Spanned<Pattern>, SpanExpr<'a>)]),
+    Loop(ExprList<'a>),
-    Func(&'a [(Spanned<Pattern>, Spanned<Type>)], Spanned<Type>, ExprRef<'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 {
+pub enum Switcher {
-    Ident(Ident),
     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,
+    Equ,
-    VLine,
+    Neq,
     Lt,
     Gt,
-    Equ,
+    LtEq,
-    Nequ,
+    GtEq,
-    LtEqu,
-    GtEqu,
 }
 
 #[derive(Clone, Copy, Debug, PartialEq, Eq)]
-pub enum UnaryOperator {
+pub enum UnaryOp {
-    Tilde,
     Minus,
+    Not,
     Star,
 }

src/syntax/parser.rs

@@ -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},
+        ast::{Definition, Ident, SpanExpr, Spanned},
-        token::Token,
+        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| {
+    select!(Token::Ident(id) => id).map_with_span(Spanned::new)
-        let ident = select!(Token::Ident(id) => id);
-
-        let literal = select! {
-            Token::Int(a) => Literal::Integer(a),
-            Token::String(a) => Literal::String(a)
-        };
-
-        let pattern = select! {
-            Token::Ident(id) => Pattern::Ident(id),
-            Token::Underscore => Pattern::None,
 }
-        .or(literal.map(Pattern::Literal))
-        .map_with_span(Spanned::new);
 
-        let type_ = just([Token::LeftParen, Token::RightParen])
+fn ty<'a, I>() -> impl Parser<'a, I, Spanned<Type>, Extra<'a>> + Clone + Copy
-            .to(Type::Unit)
+where
-            .or(ident.map(Type::Ident))
+    I: Input<'a, Token = Token, Span = SimpleSpan> + ValueInput<'a>,
-            .map_with_span(Spanned::new);
+{
+    ident().map(|i| i.map(Type::Ident))
+}
 
-        let block = expr
+fn definition<'a, I>() -> impl Parser<'a, I, Spanned<Definition<'a>>, Extra<'a>> + Clone
-            .clone()
+where
-            .separated_by(just(Token::Semicolon))
+    I: Input<'a, Token = Token, Span = SimpleSpan> + ValueInput<'a>,
-            .allow_trailing()
+{
-            .pipe(arena_collect)
+    let ident = ident();
-            .delimited_by(just(Token::LeftCurly), just(Token::RightCurly));
+    let ty = ty();
 
-        let func = just(Token::Func)
+    let func = just(T!["func"])
-            .ignore_then(
+        .ignore_then(ident)
-                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)
+            ident
-                    .ignore_then(expr.clone())
+                .then_ignore(just(T![":"]))
-                    .or(block.clone().map(Expr::Block).map_with_span(Spanned::new)),
+                .then(ty)
-            )
+                .separated_by(just(T![","]))
-            .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))
                 .allow_trailing()
                 .pipe(arena_collect)
-                .delimited_by(just(Token::LeftParen), just(Token::RightParen))
+                .delimited_by(just(T!["("]), just(T![")"])),
-                .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),
         )
+        .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
+    let binding = equivmap!(Token, DefKind, [Const, Var])
-            .clone()
+        .then(ident)
-            .map_with_state(|item, _, state| bumpalo::vec![in state.arena; item])
+        .then(just(T![":"]).ignore_then(ty).or_not())
-            .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),
-                )
-            },
-        );
-
-        // <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
+            just(T!["="]).ignore_then(
-                    .then_ignore(just(Token::RArrow))
+                just(T!["uninit"])
-                    .then(expr)
+                    .to(Expr::Uninit)
-                    .separated_by(just(Token::Comma))
+                    .map_with_span(Spanned::new),
-                    .allow_trailing()
+            ),
-                    .pipe(arena_collect)
-                    .delimited_by(just(Token::LeftCurly), just(Token::RightCurly)),
         )
-            .map_with_state(|(expr, branches), _, state| {
+        .map_with_state(
-                Expr::Match(state.arena.alloc(expr), branches)
+            |(((kind, ident), ty), init), _, state| Definition::Binding {
-            });
+                kind,
+                ident,
+                ty,
+                init: state.arena.alloc(init),
+            },
+        );
 
-        bind.or(set)
+    func.or(binding).map_with_span(Spanned::new)
-            .or(match_)
-            .or(block.map(Expr::Block))
-            .map_with_span(Spanned::new)
-            .or(binary)
-            .or(atom)
-    })
 }
 
 pub struct State<'a> {
@@ -227,8 +102,8 @@ pub fn parse_input<'a>(
 ) -> ParseResult {
     println!(
         "{:?}",
-        expr()
+        definition()
-            .separated_by(just(Token::Semicolon))
+            .separated_by(just(T![";"]))
             .allow_trailing()
             .pipe(arena_collect)
             .parse_with_state(input, &mut State { arena })

src/syntax/token.rs

@@ -1,112 +1,139 @@
-use lasso::Spur;
+use lasso::{Rodeo, Spur};
-use logos::Lexer;
+use logos::{Lexer, Logos};
-
-use {lasso::Rodeo, logos::Logos};
 
 #[derive(Default)]
-pub struct Lextras {
+pub struct Extras {
     pub interner: Rodeo,
 }
 
-#[derive(Copy, Clone, Debug, PartialEq, Eq)]
+macro_rules! token_def {
-pub enum IntLit {
+    (
-    Signed(i64),
+        unit    { $($u_name:ident : $($u_tok:literal),* $(,)?;)* }
-    Unsigned(u64),
+        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)*
             }
 
-#[derive(Logos, Copy, Clone, Debug, PartialEq, Eq)]
+            macro_rules! T {
-#[logos(extras = Lextras)]
+                $($(
-#[logos(skip r"[ \t\n\f]+")]
+                    ($u_tok) => { $crate::syntax::token::Token::$u_name };
-#[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("←")]  //____
+                $((~($kw)) => { $crate::syntax::token::Token::[<$kw:camel>] };)*
-    #[token("<-")] LArrow,
+            }
-    #[token("→")]  //____
+        });
-    #[token("->")] RArrow,
+    };
+}
 
-    #[token(":>")] Pipe,
+token_def!(
+    unit {
+        LeftParen : "(";
+        RightParen: ")";
+        LeftCurly : "{";
+        RightCurly: "}";
+        Dot       : ".";
+        Comma     : ",";
+        Colon     : ":";
+        Semicolon : ";";
 
-    #[token("+")] Plus,
+        LArrow: "←", "<-";
-    #[token("-")] Minus,
+        RArrow: "→", "->";
-    #[token("*")] Star,
-    #[token("/")] Slash,
-    #[token("&")] And,
-    #[token("|")] VLine,
-    #[token("~")] Tilde,
 
-    #[token("<")] Lt,
+        Plus   : "+";
-    #[token(">")] Gt,
+        Minus  : "-";
-    #[token("=")] Equ,
+        Star   : "*";
-    #[token("≠") ] //__
+        Slash  : "/";
-    #[token("/=")] Nequ,
+        Precent: "%";
-    #[token("≤") ] //___
-    #[token("<=")] LtEqu,
-    #[token("≥") ] //___,
-    #[token(">=")] GtEqu,
 
-    #[token("match")] Match,
+        Equ  : "=";
-    #[token("else")] Else,
+        Neq  : "≠", "/=";
-    #[token("loop")] Loop,
+        Lt   : "<";
-    #[token("const")] Const,
+        Gt   : ">";
-    #[token("var")] Var,
+        LtEq : "≤", "<=";
-    #[token("func")] Func,
+        GtEq : "≥", ">=";
-    // Modules aren't real here ondra just variables with imported functions
+    }
-    #[token("module")] Module,
 
+    keyword { func, var, const, include, switch, loop,
+              return, break, continue, uninit, asm }
+
+    else {
         #[regex(
             r"\p{XID_Start}\p{XID_Continue}*",
-        |l| l.extras.interner.get_or_intern(l.slice())
+            |l| intern(l, l.slice()),
         )] Ident(Spur),
 
-    #[token("»", better_string)]
         #[regex(
             "\"[^\"]*\"",
             |l| {
-            let slice = l.slice();
+                let s = l.slice();
-            l.extras.interner.get_or_intern(&slice[1..slice.len() - 1])
+                intern(l, &s[1..s.len() - 1])
-        }
+            },
         )] String(Spur),
 
         #[regex(
-        "-?[0-9]+",
+            "[0-9]+",
-        |l| {
+            |l| l.slice().parse::<u64>().ok()
-            Some(if let Some(slice) = l.slice().strip_prefix('-') {
+        )] Int(u64),
-                IntLit::Signed(slice.parse::<i64>().ok()?)
-            } else {
-                IntLit::Unsigned(l.slice().parse::<u64>().ok()?)
-            })
-        }
-    )] Int(IntLit),
 
         Invalid,
     }
+);
 
-// For Evy, with love.
+pub(crate) use T;
-fn better_string(lexer: &mut Lexer<Token>) -> Option<Spur> {
+
-    let mut count = 1;
+fn intern(lexer: &mut Lexer<'_, Token>, s: &str) -> Spur {
-    for (ix, chr) in lexer.remainder().char_indices() {
+    lexer.extras.interner.get_or_intern(s)
-        match chr {
-            '«' => count -= 1,
-            '»' => count += 1,
-            _ => (),
 }
 
-        if count == 0 {
+#[cfg(test)]
-            let slice = &lexer.remainder()[..ix];
+mod tests {
-            lexer.bump(ix + '«'.len_utf8());
+    use super::*;
-            return Some(lexer.extras.interner.get_or_intern(slice));
+
+    #[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
-}