From 6d866f8012df40903fa0bb664b71d899cc7e1be3 Mon Sep 17 00:00:00 2001
From: tofpie <75836434+tofpie@users.noreply.github.com>
Date: Fri, 11 Dec 2020 16:18:12 +0000
Subject: [PATCH] Fix Unicode character escape sequence parsing (#959)

Co-authored-by: tofpie <tofpie@users.noreply.github.com>
---
 boa/src/syntax/lexer/string.rs |  4 ---
 boa/src/syntax/lexer/tests.rs  | 47 ++++++++++++++++++++++++++++------
 2 files changed, 39 insertions(+), 12 deletions(-)

diff --git a/boa/src/syntax/lexer/string.rs b/boa/src/syntax/lexer/string.rs
index ad85c40e10..ab07cc9552 100644
--- a/boa/src/syntax/lexer/string.rs
+++ b/boa/src/syntax/lexer/string.rs
@@ -111,14 +111,10 @@ impl<R> Tokenizer<R> for StringLiteral {
                             b'u' => {
                                 // Support \u{X..X} (Unicode Codepoint)
                                 if cursor.next_is(b'{')? {
-                                    cursor.next_byte()?.expect("{ character vanished"); // Consume the '{'.
-
                                     // TODO: use bytes for a bit better performance (using stack)
                                     let mut code_point_buf = Vec::with_capacity(6);
                                     cursor.take_until(b'}', &mut code_point_buf)?;
 
-                                    cursor.next_byte()?.expect("} character vanished"); // Consume the '}'.
-
                                     let code_point_str = unsafe {
                                         str::from_utf8_unchecked(code_point_buf.as_slice())
                                     };
diff --git a/boa/src/syntax/lexer/tests.rs b/boa/src/syntax/lexer/tests.rs
index 4dd22e194c..accc078c9d 100644
--- a/boa/src/syntax/lexer/tests.rs
+++ b/boa/src/syntax/lexer/tests.rs
@@ -259,7 +259,7 @@ fn check_positions() {
 
 #[test]
 fn check_positions_codepoint() {
-    let s = r#"console.log("hello world\u{{2764}}"); // Test"#;
+    let s = r#"console.log("hello world\u{2764}"); // Test"#;
     // --------123456789
     let mut lexer = Lexer::new(s.as_bytes());
 
@@ -281,19 +281,19 @@ fn check_positions_codepoint() {
     // String token starts on column 13
     assert_eq!(
         lexer.next().unwrap().unwrap().span(),
-        span((1, 13), (1, 36))
+        span((1, 13), (1, 34))
     );
 
-    // Close parenthesis token starts on column 36
+    // Close parenthesis token starts on column 34
     assert_eq!(
         lexer.next().unwrap().unwrap().span(),
-        span((1, 36), (1, 37))
+        span((1, 34), (1, 35))
     );
 
-    // Semi Colon token starts on column 37
+    // Semi Colon token starts on column 35
     assert_eq!(
         lexer.next().unwrap().unwrap().span(),
-        span((1, 37), (1, 38))
+        span((1, 35), (1, 36))
     );
 }
 
@@ -702,10 +702,10 @@ fn codepoint_with_no_braces() {
 fn illegal_code_point_following_numeric_literal() {
     // Checks as per https://tc39.es/ecma262/#sec-literals-numeric-literals that a NumericLiteral cannot
     // be immediately followed by an IdentifierStart where the IdentifierStart
-    let mut lexer = Lexer::new(&br#"17.4\u{{2764}}"#[..]);
+    let mut lexer = Lexer::new(&br#"17.4\u{2764}"#[..]);
     assert!(
         lexer.next().is_err(),
-        "IdentifierStart \\u{{2764}} following NumericLiteral not rejected as expected"
+        "IdentifierStart \\u{2764} following NumericLiteral not rejected as expected"
     );
 }
 
@@ -723,6 +723,37 @@ fn non_english_str() {
     expect_tokens(&mut lexer, &expected);
 }
 
+#[test]
+fn unicode_escape_with_braces() {
+    let mut lexer = Lexer::new(&br#"'{\u{20ac}\u{a0}\u{a0}}'"#[..]);
+
+    let expected = [TokenKind::StringLiteral("{\u{20ac}\u{a0}\u{a0}}".into())];
+
+    expect_tokens(&mut lexer, &expected);
+
+    lexer = Lexer::new(&br#"\u{{a0}"#[..]);
+
+    if let Error::Syntax(_, pos) = lexer
+        .next()
+        .expect_err("Malformed Unicode character sequence expected")
+    {
+        assert_eq!(pos, Position::new(1, 1));
+    } else {
+        panic!("invalid error type");
+    }
+
+    lexer = Lexer::new(&br#"\u{{a0}}"#[..]);
+
+    if let Error::Syntax(_, pos) = lexer
+        .next()
+        .expect_err("Malformed Unicode character sequence expected")
+    {
+        assert_eq!(pos, Position::new(1, 1));
+    } else {
+        panic!("invalid error type");
+    }
+}
+
 mod carriage_return {
     use super::*;