From 32d7292e47fc8121c056715d5f995da315c905ca Mon Sep 17 00:00:00 2001
From: multipleof4 <planetrenox@protonmail.com>
Date: Fri, 26 Sep 2025 01:53:11 -0700
Subject: [PATCH] Feat: Implement robust AST parser for Hi language

---
 grammar.ne | 101 +++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 74 insertions(+), 27 deletions(-)

diff --git a/grammar.ne b/grammar.ne
index 9bd9c41..12e2906 100644
--- a/grammar.ne
+++ b/grammar.ne
@@ -5,75 +5,122 @@ const lexer = moo.compile({
   ws:      /[ \t]+/,
   nl:      { match: /\n/, lineBreaks: true },
   comment: /\/\/.*?$/,
-  number:  /0|[1-9][0-9]*/,
+  number:  /0|[1-9][0-9]*(\.[0-9]+)?/,
   string:  /"(?:\\["\\]|[^\n"\\])*"/,
   lbrace:  '{',
   rbrace:  '}',
-  identifier: /[a-zA-Z_][a-zA-Z0-9_]*/,
-  colon:   ':',
-  eq:      '=',
   lparen:  '(',
   rparen:  ')',
+  dot:     '.',
+  plus:    '+',
+  comma:   ',',
+  colon:   ':',
+  eq:      '=',
+  identifier: /[a-zA-Z_][a-zA-Z0-9_]*/,
 });
 %}
 
 @lexer lexer
 
-# Main entry point: a program is a series of statements
 Program -> _ Statements _ {% ([,,stmts]) => ({ type: 'Program', body: stmts }) %}
 
-# Statements are separated by newlines
-Statements -> Statement (_NL Statement):* _ {% 
+Statements -> Statement (_NL Statement):* _ {%
   (d) => {
     const stmts = [d[0]];
     for (const rest of d[1]) {
       stmts.push(rest[1]);
     }
-    return stmts.filter(s => s !== null); // Filter out empty lines
+    return stmts.filter(s => s !== null && s.type !== 'Comment');
   }
 %}
 
-# A statement can be a declaration, assignment, or expression
 Statement -> Declaration {% id %}
            | Assignment {% id %}
-           | OutputCall {% id %}
+           | ExpressionStatement {% id %}
            | Comment {% id %}
 
+ExpressionStatement -> Expression {% (d) => ({ type: 'ExpressionStatement', expression: d[0] }) %}
+
 Comment -> %comment {% d => ({ type: 'Comment', value: d[0].value }) %}
 
-Declaration -> %identifier _ %colon _ Expression {% 
-  (d) => ({ 
+Declaration -> %identifier _ %colon _ Expression {%
+  (d) => ({
     type: 'VariableDeclaration',
     identifier: d[0].value,
     value: d[4]
-  }) 
+  })
 %}
 
-Assignment -> %identifier _ %eq _ Expression {% 
-  (d) => ({ 
+Assignment -> %identifier _ %eq _ Expression {%
+  (d) => ({
     type: 'Assignment',
     identifier: d[0].value,
     value: d[4]
-  }) 
+  })
 %}
 
-OutputCall -> %identifier %lparen _ Expression _ %rparen {%
-  (d) => {
-    // For now, only allow '_' as the function name
-    if (d[0].value !== '_') {
-        throw new Error("Only the '_' function is supported for output.");
+# Expression parsing hierarchy to handle operator precedence
+Expression -> Additive {% id %}
+
+Additive -> Additive _ %plus _ Call {% (d) => ({ type: 'BinaryExpression', operator: '+', left: d[0], right: d[4] }) %}
+          | Call {% id %}
+
+Call -> Member ( %lparen _ (ArgumentList):? _ %rparen ):? {%
+    (d) => {
+        if (d[1]) { // It is a function call
+            return {
+                type: 'CallExpression',
+                callee: d[0],
+                arguments: d[1][2] || []
+            }
+        }
+        return d[0]; // Not a call, just a Member/Primary expression
     }
-    return { type: 'OutputCall', arguments: [d[3]] }
-  }
 %}
 
-# Expressions (very simple for now)
-Expression -> Literal {% id %}
+Member -> Primary ( %dot %identifier ):* {%
+    (d) => {
+        let obj = d[0];
+        for (const part of d[1]) {
+            obj = {
+                type: 'MemberExpression',
+                object: obj,
+                property: { type: 'Identifier', name: part[1].value }
+            }
+        }
+        return obj;
+    }
+%}
+
+Primary -> Literal {% id %}
+         | Block {% id %}
+         | Identifier {% id %}
+         | %lparen _ Expression _ %rparen {% (d) => d[2] %}
+
+ArgumentList -> Expression ( _ %comma _ Expression):* {%
+    (d) => [d[0], ...d[1].map(m => m[3])]
+%}
+
+Identifier -> %identifier {% d => ({ type: 'Identifier', name: d[0].value }) %}
 
 Literal -> %number {% d => ({ type: 'NumericLiteral', value: Number(d[0].value) }) %}
          | %string {% d => ({ type: 'StringLiteral', value: d[0].value }) %}
 
-# Whitespace and Newlines
+Block -> %lbrace _ (KeyValuePairs):? _ %rbrace {%
+  (d) => ({ type: 'Block', properties: d[2] || [] })
+%}
+
+KeyValuePairs -> KeyValuePair (_NL KeyValuePair):* {%
+  (d) => [d[0], ...d[1].map(m => m[1])]
+%}
+
+KeyValuePair -> %identifier _ %colon _ Expression {%
+  (d) => ({
+    type: 'Property',
+    key: d[0].value,
+    value: d[4]
+  })
+%}
+
 _ -> (%ws | %comment):* {% () => null %}
 _NL -> (_ %nl):+ {% () => null %}
-