From c98550907d00bf5e3c7c24f71ed08dae53887a27 Mon Sep 17 00:00:00 2001
From: Nixon Enraght-Moony <nixon.emoony@gmail.com>
Date: Wed, 29 Dec 2021 16:49:49 +0000
Subject: [PATCH] Basic parser

---
 .gitignore    |   5 +++
 Cargo.toml    |  11 ++++++
 hello.scm     |  11 ++++++
 src/main.rs   |  45 +++++++++++++++++++++
 src/parser.rs | 107 ++++++++++++++++++++++++++++++++++++++++++++++++++
 5 files changed, 179 insertions(+)
 create mode 100644 Cargo.toml
 create mode 100644 hello.scm
 create mode 100644 src/main.rs
 create mode 100644 src/parser.rs

diff --git a/.gitignore b/.gitignore
index f5503f1..8c95310 100644
--- a/.gitignore
+++ b/.gitignore
@@ -23,3 +23,8 @@ Cargo.lock
 *.scm#*
 .#*.scm
 
+
+
+# Added by cargo
+
+/target
diff --git a/Cargo.toml b/Cargo.toml
new file mode 100644
index 0000000..dee7742
--- /dev/null
+++ b/Cargo.toml
@@ -0,0 +1,11 @@
+[package]
+name = "handball"
+version = "0.1.0"
+edition = "2021"
+
+[features]
+
+[dependencies]
+anyhow = "1.0.52"
+fs-err = "2.6.0"
+logos = "0.12.0"
diff --git a/hello.scm b/hello.scm
new file mode 100644
index 0000000..11d8dd0
--- /dev/null
+++ b/hello.scm
@@ -0,0 +1,11 @@
+; https://github.com/munificent/craftinginterpreters/blob/master/test/closure/closed_closure_in_function.lox
+
+(define f #f)
+
+((lambda ()
+   (define local 1)
+   (define (f_)
+     (display local))
+   (set! f f_)))
+
+(f)
diff --git a/src/main.rs b/src/main.rs
new file mode 100644
index 0000000..9b78de7
--- /dev/null
+++ b/src/main.rs
@@ -0,0 +1,45 @@
+mod parser;
+
+use fs_err as fs;
+
+#[derive(Debug, PartialEq, logos::Logos)]
+enum Kind {
+    #[token("(")]
+    Lparen,
+
+    #[token(")")]
+    Rparen,
+
+    #[regex(r"[A-Za-z0-9!$%&*+-./:<=>?@^_~]+")]
+    // Most literals are floats, and can be seperated later
+    Symbol,
+
+    #[regex(r"#(f|t)")]
+    Boolean,
+
+    #[regex(r#""(\\.|[^"\\])*""#)]
+    String,
+
+    #[error]
+    #[regex(r"[ \t\n\f\r]+", logos::skip)]
+    #[regex(r";.*", logos::skip)]
+    Error,
+}
+
+fn main() -> anyhow::Result<()> {
+    let arg = std::env::args()
+        .nth(1)
+        .ok_or_else(|| anyhow::anyhow!("Useage: handball sources.scm"))?;
+
+    let src = fs::read_to_string(&arg)?;
+
+    dbg!(parser::parse(&src));
+
+    Ok(())
+}
+
+#[derive(Debug)]
+enum Ast<'s> {
+    Tree(Vec<Ast<'s>>),
+    Leaf(&'s str),
+}
diff --git a/src/parser.rs b/src/parser.rs
new file mode 100644
index 0000000..7fae83d
--- /dev/null
+++ b/src/parser.rs
@@ -0,0 +1,107 @@
+use std::iter::Peekable;
+
+use logos::Logos;
+
+use crate::{Ast, Kind};
+
+type Error = String;
+
+struct Parser<'s> {
+    lex: Peekable<Lexer<'s>>,
+    errors: Vec<Error>,
+}
+
+pub(crate) fn parse(s: &str) -> (Vec<Ast>, Vec<Error>) {
+    let lex = Lexer {
+        lex: Kind::lexer(s),
+    }
+    .peekable();
+
+    let mut parser = Parser {
+        lex,
+        errors: vec![],
+    };
+
+    let mut trees = Vec::new();
+
+    while parser.lex.peek().is_some() {
+        if let Some(t) = parser.parse() {
+            trees.push(t);
+        }
+    }
+
+    (trees, parser.errors)
+}
+
+impl<'s> Parser<'s> {
+    fn parse(&mut self) -> Option<Ast<'s>> {
+        let Token { kind, span, src } = self.next()?;
+
+        // match tok {
+        //     Token::Lparen => todo!(),
+        //     Token::Rparen => {
+        //         self.errors.push(format!("Unexpected lparen "));
+        //         self.parse()
+        //     }
+        //     Token::Symbol => Some(Ast::Leaf(self.lex.str())),
+        //     Token::String => todo!(),
+        //     Token::Error => todo!(),
+        // }
+        Some(match kind {
+            Kind::Lparen => {
+                let mut bits = Vec::new();
+
+                while self.lex.peek()?.kind != Kind::Rparen {
+                    bits.push(self.parse()?);
+                }
+
+                let rp = self.next()?;
+                assert_eq!(rp.kind, Kind::Rparen);
+
+                Ast::Tree(bits)
+            }
+            Kind::Rparen => {
+                self.errors.push(format!("Unexpected `)` at {:?}", span));
+                // Keep going from the next token
+                return self.parse();
+            }
+            Kind::Symbol | Kind::String | Kind::Boolean => Ast::Leaf(src),
+            Kind::Error => unreachable!("Removed by Parser::next"),
+        })
+    }
+
+    fn next(&mut self) -> Option<Token<'s>> {
+        for tok in &mut self.lex {
+            if tok.kind == crate::Kind::Error {
+                self.errors
+                    .push(format!("Bad token at {:?}: `{}`", tok.span, tok.src))
+            } else {
+                return Some(tok);
+            }
+        }
+        self.errors.push("Unexpected EOF".to_owned());
+        None
+    }
+}
+
+struct Lexer<'s> {
+    lex: logos::Lexer<'s, crate::Kind>,
+}
+
+impl<'s> Iterator for Lexer<'s> {
+    type Item = Token<'s>;
+
+    fn next(&mut self) -> Option<Self::Item> {
+        self.lex.next().map(|kind| Token {
+            kind,
+            span: self.lex.span(),
+            src: self.lex.slice(),
+        })
+    }
+}
+
+struct Token<'s> {
+    kind: Kind,
+    span: logos::Span,
+    src: &'s str,
+}