first part of static type system

src/main.rs

@@ -15,7 +15,7 @@ fn main() {
 r"
 pi = 3.1415926535
 
-sin(x) = {
+sin(x: f4) = {
     x
 }
 

src/parser/data.rs

@@ -1,12 +1,13 @@
 use std::collections::{VecDeque};
 use crate::token::{Token};
+use crate::Prim;
 
-#[derive(Eq, Debug)]
+#[derive(Debug)]
 pub struct Func<'a> {
     /// name of this function
     pub name: Option<&'a str>,
     /// parameter names
-    pub args: Option<Vec<&'a str>>,
+    pub args: Option<Vec<(&'a str, Prim)>>,
     /// raw tokens
     pub raw: Option<VecDeque<Token<'a>>>,
     /// if the function returns a single value
@@ -43,11 +44,11 @@ impl<'a> std::fmt::Display for Func<'a> {
             f.write_str("(")?;
             for (x, arg) in args.iter().enumerate() {
                 if x == 0 {
-                    f.write_fmt(format_args!("{}", arg))?;
+                    f.write_fmt(format_args!("{}", arg.0))?;
                     continue;
                 }
 
-                f.write_fmt(format_args!(", {}", arg))?;
+                f.write_fmt(format_args!(", {}", arg.0))?;
             }
             f.write_str(")")?;
         }
@@ -62,7 +63,7 @@ impl<'a> std::fmt::Display for Func<'a> {
 
 pub type Block<'a> = VecDeque<Expr<'a>>;
 
-#[derive(Debug)]
+#[derive(Debug, Clone)]
 pub enum Expr<'a> {
     /// group of more expressions
     Block(Block<'a>),
@@ -72,9 +73,9 @@ pub enum Expr<'a> {
 
 pub struct Scope<'a> {
     pub funcs: Vec<&'a str>,
-    pub args: Option<&'a Vec<&'a str>>,
+    pub args: Option<&'a Vec<(&'a str, Prim)>>,
     /// stack of scoped block variables
-    pub vars: Vec<Vec<String>>,
+    pub vars: Vec<Vec<(String, Option<Prim>)>>,
 }
 
 impl<'a> Scope<'a> {
@@ -86,8 +87,8 @@ impl<'a> Scope<'a> {
         self.vars.pop();
     }
 
-    pub fn decl_var(&mut self, name: String) {
+    pub fn decl_var(&mut self, name: String, typ: Option<Prim>) {
-        self.vars.last_mut().unwrap().push(name)
+        self.vars.last_mut().unwrap().push((name, typ))
     }
 
     pub fn is_func(&self, name: &'a str) -> bool {
@@ -96,21 +97,27 @@ impl<'a> Scope<'a> {
     
     pub fn is_arg(&self, name: &'a str) -> bool {
         if let Some(args) = self.args {
-            return args.contains(&name);
+            for arg in args.iter() {
+                if arg.0 == name {
+                    return true;
+                }
+            }
         }
         false
     }
 
-    pub fn is_var(&self, name: &'a str) -> bool {
+    pub fn is_var(&self, name: &'a str) -> Option<Prim> {
         // create an owned version of the string
         let owned = &name.to_owned();
         
         // search
         for vars in self.vars.iter() {
-            if vars.contains(owned) {
+            for var in vars.iter() {
-                return true;
+                if &var.0 == owned {
+                    return var.1;
                 }
             }
-        false
+        }
+        None
     }
 }

src/parser/mod.rs

@@ -1,6 +1,6 @@
-use core::panic;
+use core::{panic};
 use std::{collections::{VecDeque}, vec};
-use crate::token::{Token, Operator, Assoc};
+use crate::token::{Token, Operator, Assoc, Prim};
 
 pub mod data;
 
@@ -86,7 +86,7 @@ fn discover_functions<'a>(tokens: &mut VecDeque<crate::Token<'a>>) -> Vec<Func<'
                     _ => ()
                 }
 
-                Token::Assign(name) => {
+                Token::Assign(name, _) => {
                     if func.results {
                         panic!("double function assignment not permitted")
                     }
@@ -151,8 +151,9 @@ fn discover_functions<'a>(tokens: &mut VecDeque<crate::Token<'a>>) -> Vec<Func<'
             }
 
             match &top {
-                Token::Word(text) => args.push(text),
+                Token::Decl(name, typ) => args.push((name, *typ)),
-                _ => panic!("Argument is not a word {:?}", &top)
+                Token::Word(name) => panic!("Missing type declaration {name}"),
+                _ => panic!("Argument is not a declaration {:?}", &top)
             }
             continue;
         }
@@ -221,6 +222,14 @@ fn discover_exprs<'a>(functions: &mut Vec<Func<'a>>) {
 fn parse_term<'a>(term: &mut VecDeque<Token<'a>>, scope: &mut Scope) {
     let mut op_stack = vec![];
     let mut output = VecDeque::with_capacity(term.len());
+    let mut value_stack = vec![];
+
+    /*
+    
+    Token::Number(text) => value_stack.push(CompileTimeType::UntypedNum(text)),
+    Token::Bool(_) => value_stack.push(CompileTimeType::Prim(Prim::Bool)),
+    
+    */
 
     'outer:
     while let Some(token) = term.pop_front() {
@@ -232,15 +241,18 @@ fn parse_term<'a>(term: &mut VecDeque<Token<'a>>, scope: &mut Scope) {
                 } else if scope.is_arg(text) {
                     output.push_back(Token::Arg(text));
                     continue;
-                } else if scope.is_var(text) {
+                } else if scope.is_var(text).is_some() {
                     output.push_back(Token::Var(text));
                     continue;
                 }
                 panic!("Unknwon word: {text}")
             }
-            Token::Number(_) => output.push_back(token),
+            Token::Number(_) => {
-            Token::Assign(text) => {
+                output.push_back(token);
-                scope.decl_var((*text).to_owned());
+                value_stack.push(CompileTimeType::UntypedNum)
+            },
+            Token::Assign(text, typ) => {
+                scope.decl_var((*text).to_owned(), typ.to_owned());
                 op_stack.push(token);
             },
             Token::Keyword(_) => op_stack.push(token),
@@ -303,6 +315,11 @@ fn parse_term<'a>(term: &mut VecDeque<Token<'a>>, scope: &mut Scope) {
     term.append(&mut output);
 }
 
+enum CompileTimeType {
+    Prim(Prim),
+    UntypedNum,
+}
+
 fn parse_block(block: &mut Block, scope: &mut Scope) {
     scope.alloc_scope();
     for expr in block.iter_mut() {

src/token/mod.rs

@@ -57,13 +57,7 @@ impl Operator {
             Operator::Lt => 2,
             Operator::Gt => 2,
             Operator::LtEq => 2,
-            Operator::GtEq => 2,       Operator::Eq => 2,
-            Operator::Lt => 2,
-            Operator::Gt => 2,
-            Operator::LtEq => 2,
             Operator::GtEq => 2,
-            Operator::NotEq => 2,
-            Operator::NotEq => 2,
 
             Operator::Or => 0,
             Operator::Xor => 0,
@@ -108,15 +102,25 @@ impl Keyword {
     }
 }
 
-pub struct SourceString<'a> {
+#[derive(Debug, PartialEq, Eq, Copy, Clone)]
-    pub string: &'a str,
+pub enum Prim {
-    /// line in which the source string is to be found
+    Int,
-    pub line: usize,
+    Real,
-    /// index in source where the token starts
+    Bool,
-    pub start: usize
 }
 
-#[derive(Debug, Hash, PartialEq, Eq, Copy, Clone)]
+impl Prim {
+    fn from<'a>(text: &'a str) -> Prim {
+        return match text {
+            "i4" => Prim::Int,
+            "f4" => Prim::Real,
+            "bool" => Prim::Bool,
+            _ => panic!("Unknown type declaration: {text}")
+        }
+    }
+}
+
+#[derive(Debug, PartialEq, Eq, Copy, Clone)]
 /// A token represents a basic building block for source code.
 /// They give a meaning to patterns of chars allowing to interpret them.
 pub enum Token<'a> {
@@ -129,13 +133,13 @@ pub enum Token<'a> {
     Func(&'a str),
     Var(&'a str),
     Arg(&'a str),
-    Assign(&'a str),
+    Assign(&'a str, Option<Prim>),
+    Decl(&'a str, Prim),
     Bool(bool),
     Keyword(Keyword),
-    TypeDecl(&'a str)
 }
 
-const TOKEN_REGEX_SRC: &'static str = r"(#.*)|(if|while|loop|break|continue)|(true|false|yes|no|maybe)|([A-Za-z_]+)\s*(?::\s*(i4|f4|bool))?\s*=|([A-Za-z_]+)|(\d*\.?\d+)|(!=|==|<=|<=|[&|+\-*/<>])|([(){}])|(\n)";
+const TOKEN_REGEX_SRC: &'static str = r"(#.*)|(if|while|loop|break|continue)|(true|false|yes|no|maybe)|([A-Za-z_]+)\s*(?::\s*([a-zA-Z0-9]+))|([A-Za-z_]+)\s*(?::\s*([a-zA-Z0-9]+))?\s*=|([A-Za-z_]+)|(\d*\.?\d+)|(!=|==|<=|<=|[&|+\-*/<>])|([(){}])|(\n+)";
  
 lazy_static::lazy_static! {
     static ref TOKEN_REGEX: regex::Regex = regex::Regex::new(TOKEN_REGEX_SRC).unwrap();
@@ -146,7 +150,14 @@ pub fn tokenize<'a>(source: &'a str) -> VecDeque<Token<'a>> {
     let mut tokens = VecDeque::new();
 
     for cap in TOKEN_REGEX.captures_iter(source) {
-        for (i, group) in cap.iter().enumerate() {
+        let mut enumerator = cap.iter().enumerate();
+        loop {
+            let next = enumerator.next();
+            if next.is_none() {
+                break
+            }
+
+            let (i, group) = next.unwrap();
 
             // ignore first group as its the entire match,
             // as well as the 1st group (= comments)
@@ -159,14 +170,25 @@ pub fn tokenize<'a>(source: &'a str) -> VecDeque<Token<'a>> {
                 tokens.push_back(match i {
                     2 => Token::Keyword(Keyword::parse(mat.as_str())),
                     3 => Token::Bool(parse_bool(mat.as_str())),
-                    4 => Token::Assign(mat.as_str()),
+                    4 => {
-                    5 => Token::Word(mat.as_str()),
+                        let var_type = Prim::from(enumerator.next().unwrap().1.unwrap().as_str());
-                    6 => Token::Number(mat.as_str()),
+                        Token::Decl(mat.as_str(), var_type)
-                    7 => Token::Operator(Operator::parse(mat.as_str())),
+                    },
-                    8 => Token::Delemiter(mat.as_str().chars().nth(0).unwrap()),
+                    6 => {
-                    9 => Token::LineBreak,
+                        let var_type = if let Some(mat) = enumerator.next().unwrap().1 {
+                            Some(Prim::from(mat.as_str()))
+                        } else {
+                            None
+                        };
+                        Token::Assign(mat.as_str(), var_type)
+                    },
+                    8 => Token::Word(mat.as_str()),
+                    9 => Token::Number(mat.as_str()),
+                    10 => Token::Operator(Operator::parse(mat.as_str())),
+                    11 => Token::Delemiter(mat.as_str().chars().nth(0).unwrap()),
+                    12 => Token::LineBreak,
 
-                    _ => panic!("Unknown match to tokenize: {}", mat.as_str())
+                    _ => panic!("Unknown match to tokenize ({i}): {}", mat.as_str())
                 });
                 break;
             }