allow commas in numbers

engine/src/evaluator.zig

@@ -486,6 +486,16 @@ test "eval number with underscores" {
     try testing.expectEqual(@as(f64, 1_000_001.0), result);
 }
 
+test "eval number with commas" {
+    const result = try testEval("1,000 * 2.3");
+    try testing.expectApproxEqAbs(@as(f64, 2300.0), result, 1e-10);
+}
+
+test "eval commas not confused with function args" {
+    const result = try testEval("max(1,000, 500)");
+    try testing.expectEqual(@as(f64, 1000.0), result);
+}
+
 test "eval bitwise not in standard mode" {
     const result = try testEval("~0");
     // ~0 as i64 = -1

engine/src/tokenizer.zig

@@ -93,7 +93,7 @@ pub fn parseNumber(token_text: []const u8) !NumberValue {
     var buf: [128]u8 = undefined;
     var buf_len: usize = 0;
     for (token_text) |c| {
-        if (c != '_') {
+        if (c != '_' and c != ',') {
             if (buf_len >= buf.len) return error.InvalidNumber;
             buf[buf_len] = c;
             buf_len += 1;
@@ -367,6 +367,14 @@ pub const Tokenizer = struct {
             } else if (ch == '_') {
                 // Digit separator
                 self.pos += 1;
+            } else if (ch == ',') {
+                // Comma as digit separator, but only if followed by a digit
+                // (to avoid eating commas in function args like max(1, 2))
+                if (self.pos + 1 < self.source.len and predicate(self.source[self.pos + 1])) {
+                    self.pos += 1;
+                } else {
+                    break;
+                }
             } else {
                 break;
             }
@@ -474,6 +482,24 @@ test "tokenize hex with underscores" {
     try testing.expectEqualStrings("0xFF_FF", t.text("0xFF_FF"));
 }
 
+test "tokenize number with commas" {
+    var tok = Tokenizer.init("1,000,000", .standard);
+    const t = tok.next();
+    try testing.expectEqual(TokenKind.number, t.kind);
+    try testing.expectEqualStrings("1,000,000", t.text("1,000,000"));
+    try testing.expectEqual(TokenKind.eof, tok.next().kind);
+}
+
+test "tokenize comma not eaten in function args" {
+    var tok = Tokenizer.init("max(1, 2)", .standard);
+    try testing.expectEqual(TokenKind.identifier, tok.next().kind); // max
+    try testing.expectEqual(TokenKind.left_paren, tok.next().kind); // (
+    try testing.expectEqual(TokenKind.number, tok.next().kind); // 1
+    try testing.expectEqual(TokenKind.comma, tok.next().kind); // ,
+    try testing.expectEqual(TokenKind.number, tok.next().kind); // 2
+    try testing.expectEqual(TokenKind.right_paren, tok.next().kind); // )
+}
+
 test "tokenize function call" {
     var tok = Tokenizer.init("sin(3.14)", .standard);
     try testing.expectEqual(TokenKind.identifier, tok.next().kind);
@@ -584,6 +610,12 @@ test "parseNumber hex with underscores" {
     try testing.expectEqual(Base.hex, result.base);
 }
 
+test "parseNumber with commas" {
+    const result = try parseNumber("1,000,000");
+    try testing.expectEqual(@as(?u64, 1_000_000), result.int_value);
+    try testing.expectEqual(Base.decimal, result.base);
+}
+
 // -- ImplicitMulStream tests --
 
 test "implicit mul: number followed by identifier (2pi)" {