Implement String.prototype.codePointAt

Closes boa-dev#751.

boa/src/builtins/date/tests.rs

@@ -1276,7 +1276,9 @@ fn date_proto_to_string() -> Result<(), Box<dyn std::error::Error>> {
 
     assert_eq!(
         Some(Value::string(
-            Local::now()
+            Local
+                .ymd(2020, 6, 8)
+                .and_hms(9, 16, 15)
                 .format("Wed Jul 08 2020 09:16:15 GMT%:z")
                 .to_string()
         )),

boa/src/builtins/string/mod.rs

@@ -50,11 +50,11 @@ pub(crate) fn code_point_at(string: RcString, position: i32) -> Option<(u32, u8,
 }
 
 fn is_leading_surrogate(value: u16) -> bool {
-    value >= 0xD800 && value <= 0xDBFF
+    (0xD800..0xDBFF).contains(&value)
 }
 
 fn is_trailing_surrogate(value: u16) -> bool {
-    value >= 0xDC00 && value <= 0xDFFF
+    (0xDC00..0xDFFF).contains(&value)
 }
 
 /// JavaScript `String` implementation.
@@ -84,6 +84,7 @@ impl BuiltIn for String {
         .property("length", 0, attribute)
         .method(Self::char_at, "charAt", 1)
         .method(Self::char_code_at, "charCodeAt", 1)
+        .method(Self::code_point_at, "codePointAt", 1)
         .method(Self::to_string, "toString", 0)
         .method(Self::concat, "concat", 1)
         .method(Self::repeat, "repeat", 1)
@@ -197,14 +198,17 @@ impl String {
             .unwrap_or_else(Value::undefined)
             .to_integer(context)? as i32;
 
+        // Fast path returning empty string when pos is obviously out of range
+        if pos < 0 || pos >= primitive_val.len() as i32 {
+            return Ok("".into());
+        }
+
         // Calling .len() on a string would give the wrong result, as they are bytes not the number of
         // unicode code points
         // Note that this is an O(N) operation (because UTF-8 is complex) while getting the number of
         // bytes is an O(1) operation.
-        let length = primitive_val.chars().count();
-
-        // We should return an empty string is pos is out of range
-        if pos >= length as i32 || pos < 0 {
+        let length = primitive_val.chars().count() as i32;
+        if pos >= length {
             return Ok("".into());
         }
 
@@ -216,6 +220,46 @@ impl String {
         ))
     }
 
+    /// `String.prototype.codePointAt( index )`
+    ///
+    /// The `codePointAt()` method returns an integer between `0` and `65535` representing the UTF-16 code unit at the given index.
+    ///
+    /// If no UTF-16 surrogate pair begins at the index, the code point at the index is returned.
+    ///
+    /// `codePointAt()` returns `undefined` if the given index is less than `0`, or if it is equal to or greater than the `length` of the string.
+    ///
+    /// More information:
+    ///  - [ECMAScript reference][spec]
+    ///  - [MDN documentation][mdn]
+    ///
+    /// [spec]: https://tc39.es/ecma262/#sec-string.prototype.codepointat
+    /// [mdn]: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/String/codePointAt
+    pub(crate) fn code_point_at(
+        this: &Value,
+        args: &[Value],
+        context: &mut Context,
+    ) -> Result<Value> {
+        // First we get it the actual string a private field stored on the object only the context has access to.
+        // Then we convert it into a Rust String by wrapping it in from_value
+        let primitive_val = this.to_string(context)?;
+        let pos = args
+            .get(0)
+            .cloned()
+            .unwrap_or_else(Value::undefined)
+            .to_integer(context)? as i32;
+
+        // Fast path returning undefined when pos is obviously out of range
+        if pos < 0 || pos >= primitive_val.len() as i32 {
+            return Ok(Value::undefined());
+        }
+
+        if let Some((code_point, _, _)) = code_point_at(primitive_val, pos) {
+            Ok(Value::from(code_point))
+        } else {
+            Ok(Value::undefined())
+        }
+    }
+
     /// `String.prototype.charCodeAt( index )`
     ///
     /// The `charCodeAt()` method returns an integer between `0` and `65535` representing the UTF-16 code unit at the given index.
@@ -238,17 +282,21 @@ impl String {
         // First we get it the actual string a private field stored on the object only the context has access to.
         // Then we convert it into a Rust String by wrapping it in from_value
         let primitive_val = this.to_string(context)?;
-
-        // Calling .len() on a string would give the wrong result, as they are bytes not the number of unicode code points
-        // Note that this is an O(N) operation (because UTF-8 is complex) while getting the number of bytes is an O(1) operation.
-        let length = primitive_val.chars().count();
         let pos = args
             .get(0)
             .cloned()
             .unwrap_or_else(Value::undefined)
             .to_integer(context)? as i32;
 
-        if pos >= length as i32 || pos < 0 {
+        // Fast path returning NaN when pos is obviously out of range
+        if pos < 0 || pos >= primitive_val.len() as i32 {
+            return Ok(Value::from(NAN));
+        }
+
+        // Calling .len() on a string would give the wrong result, as they are bytes not the number of unicode code points
+        // Note that this is an O(N) operation (because UTF-8 is complex) while getting the number of bytes is an O(1) operation.
+        let length = primitive_val.chars().count() as i32;
+        if pos >= length {
             return Ok(Value::from(NAN));
         }
 

boa/src/builtins/string/tests.rs

@@ -775,6 +775,7 @@ fn last_index_non_integer_position_argument() {
 #[test]
 fn char_at() {
     let mut context = Context::new();
+    assert_eq!(forward(&mut context, "'abc'.charAt(-1)"), "\"\"");
     assert_eq!(forward(&mut context, "'abc'.charAt(1)"), "\"b\"");
     assert_eq!(forward(&mut context, "'abc'.charAt(9)"), "\"\"");
     assert_eq!(forward(&mut context, "'abc'.charAt()"), "\"a\"");
@@ -784,12 +785,27 @@ fn char_at() {
 #[test]
 fn char_code_at() {
     let mut context = Context::new();
+    assert_eq!(forward(&mut context, "'abc'.charCodeAt(-1)"), "NaN");
     assert_eq!(forward(&mut context, "'abc'.charCodeAt(1)"), "98");
     assert_eq!(forward(&mut context, "'abc'.charCodeAt(9)"), "NaN");
     assert_eq!(forward(&mut context, "'abc'.charCodeAt()"), "97");
     assert_eq!(forward(&mut context, "'abc'.charCodeAt(null)"), "97");
 }
 
+#[test]
+fn code_point_at() {
+    let mut context = Context::new();
+    assert_eq!(forward(&mut context, "'abc'.codePointAt(-1)"), "undefined");
+    assert_eq!(forward(&mut context, "'abc'.codePointAt(1)"), "98");
+    assert_eq!(forward(&mut context, "'abc'.codePointAt(9)"), "undefined");
+    assert_eq!(forward(&mut context, "'abc'.codePointAt()"), "97");
+    assert_eq!(forward(&mut context, "'abc'.codePointAt(null)"), "97");
+    assert_eq!(
+        forward(&mut context, "'\\uD800\\uDC00'.codePointAt(0)"),
+        "65536"
+    );
+}
+
 #[test]
 fn slice() {
     let mut context = Context::new();

test.js

@@ -0,0 +1,4 @@
+console.log("☃★♲".codePointAt(1));           // 9733
+console.log('ABC'.codePointAt(1));           // 66
+console.log('\uD800\uDC00'.codePointAt(0));  // 65536
+console.log('XYZ'.codePointAt(42));          // undefined