From ad571700100c3cb8c472b681006c2b0581260ff1 Mon Sep 17 00:00:00 2001
From: Amneesh Singh <natto@weirdnatto.in>
Date: Sun, 13 Aug 2023 07:35:20 +0530
Subject: [PATCH] lexer: allow underscore for identifiers

Signed-off-by: Amneesh Singh <natto@weirdnatto.in>
---
 src/ast.rs   | 56 +++++++++++++++++++++++++++++++++++-----------------
 src/lexer.rs | 10 +++++-----
 2 files changed, 43 insertions(+), 23 deletions(-)

diff --git a/src/ast.rs b/src/ast.rs
index 81fc08a..1b17cb1 100644
--- a/src/ast.rs
+++ b/src/ast.rs
@@ -82,24 +82,6 @@ pub struct Let {
     pub expr: Option<Expr>,
 }
 
-type Op = crate::lexer::TokenSymbol;
-
-/// Lowest form of expression
-///
-/// TODO: refine
-#[derive(Debug, PartialEq)]
-pub enum Expr {
-    Int(i32),
-    Float(f32),
-    Char(char),
-    Op(Op, Box<Expr>, Option<Box<Expr>>),
-    If(Box<Expr>, Box<Expr>, Option<Box<Expr>>),
-    Block(Vec<Statement>),
-    Loop(Vec<Statement>),
-    Break,
-    Continue,
-}
-
 /// Primitives
 ///
 /// TODO: add arrays and pointers maybe
@@ -109,3 +91,41 @@ pub enum Ty {
     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>),
+}
+
+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>>),
+}
diff --git a/src/lexer.rs b/src/lexer.rs
index 6a1460c..b8695dd 100644
--- a/src/lexer.rs
+++ b/src/lexer.rs
@@ -14,7 +14,7 @@ pub enum TokenLiteral {
 }
 
 /// All token symbols
-#[derive(Debug, PartialEq)]
+#[derive(Debug, PartialEq, Clone, Copy)]
 pub enum TokenSymbol {
     // arithmetic
     Plus,
@@ -38,9 +38,7 @@ pub enum TokenSymbol {
 
     // relational
     Gt,
-    Ge,
     Lt,
-    Le,
     GtEq,
     LtEq,
     EqEq,
@@ -88,6 +86,7 @@ pub enum TokenKeyword {
     Loop,
     Break,
     Continue,
+    Return,
 
     // primitives
     Int,
@@ -280,7 +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();
@@ -302,6 +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),
@@ -409,7 +409,7 @@ impl<'a> Lexer<'a> {
                     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(),
             }