LibWeb: Add TextEncoder encodeInto

Author: https://github.com/bplaat
Commit: https://github.com/SerenityOS/serenity/commit/0104225d9b
Pull-request: https://github.com/SerenityOS/serenity/pull/21351

Meta/Lagom/Tools/CodeGenerators/LibWeb/BindingsGenerator/IDLGenerators.cpp

@@ -87,6 +87,7 @@ static bool is_javascript_builtin(Type const& type)
         "ArrayBuffer"sv,
         "Float32Array"sv,
         "Float64Array"sv,
+        "Uint8Array"sv
     };
 
     return types.span().contains_slow(type.name());
@@ -602,7 +603,7 @@ static void generate_to_cpp(SourceGenerator& generator, ParameterType& parameter
     auto @cpp_name@ = JS::make_handle(TRY(@js_name@@js_suffix@.to_object(vm)));
 )~~~");
         }
-    } else if (parameter.type->name() == "BufferSource" || parameter.type->name() == "Float32Array" || parameter.type->name() == "Float64Array") {
+    } else if (parameter.type->name() == "BufferSource" || parameter.type->name() == "Float32Array" || parameter.type->name() == "Float64Array" || parameter.type->name() == "Uint8Array") {
         if (optional) {
             scoped_generator.append(R"~~~(
     Optional<JS::Handle<JS::Object>> @cpp_name@;

Tests/LibWeb/Text/expected/Encoding/TextEncoder-encode.txt

@@ -0,0 +1,4 @@
+1. {"0":87,"1":101,"2":108,"3":108,"4":32,"5":72,"6":101,"7":108,"8":108,"9":111,"10":32,"11":70,"12":114,"13":105,"14":101,"15":110,"16":100,"17":115,"18":32,"19":240,"20":159,"21":152,"22":128,"23":33}
+2. {"0":87,"1":101,"2":108,"3":108,"4":32,"5":72,"6":101,"7":108,"8":108,"9":111,"10":32,"11":70,"12":114,"13":105,"14":101,"15":110,"16":100,"17":115,"18":32,"19":240,"20":159,"21":152,"22":128,"23":33,"24":0,"25":0,"26":0,"27":0,"28":0,"29":0,"30":0,"31":0}
+3. 22
+4. 24

Tests/LibWeb/Text/input/Encoding/TextEncoder-encode.html

@@ -0,0 +1,33 @@
+<script src="../include.js"></script>
+<script>
+    test(() => {
+        let testCounter = 1;
+        function testPart(part) {
+            println(`${testCounter++}. ${JSON.stringify(part())}`);
+        }
+
+        // 1. TextEncoder encode
+        testPart(() => new TextEncoder().encode('Well Hello Friends 😀!'));
+
+        // 2. TextEncoder encodeInto
+        testPart(() => {
+            const buffer = new Uint8Array(32);
+            new TextEncoder().encodeInto('Well Hello Friends 😀!', buffer);
+            return buffer
+        });
+
+        // 3. TextEncoder encodeInto result read
+        testPart(() => {
+            const buffer = new Uint8Array(32);
+            const result = new TextEncoder().encodeInto('Well Hello Friends 😀!', buffer);
+            return result.read;
+        });
+
+        // 3. TextEncoder encodeInto result written
+        testPart(() => {
+            const buffer = new Uint8Array(32);
+            const result = new TextEncoder().encodeInto('Well Hello Friends 😀!', buffer);
+            return result.written;
+        });
+    });
+</script>

Userland/Libraries/LibWeb/Encoding/TextEncoder.cpp

@@ -30,10 +30,9 @@ void TextEncoder::initialize(JS::Realm& realm)
 }
 
 // https://encoding.spec.whatwg.org/#dom-textencoder-encode
-JS::Uint8Array* TextEncoder::encode(String const& input) const
+JS::NonnullGCPtr<JS::Uint8Array> TextEncoder::encode(String const& input) const
 {
-    // NOTE: The AK::DeprecatedString returned from PrimitiveString::string() is always UTF-8, regardless of the internal string type, so most of these steps are no-ops.
-
+    // NOTE: The AK::String is always UTF-8, so most of these steps are no-ops.
     // 1. Convert input to an I/O queue of scalar values.
     // 2. Let output be the I/O queue of bytes « end-of-queue ».
     // 3. While true:
@@ -48,10 +47,67 @@ JS::Uint8Array* TextEncoder::encode(String const& input) const
     return JS::Uint8Array::create(realm(), array_length, *array_buffer);
 }
 
+// https://encoding.spec.whatwg.org/#dom-textencoder-encodeinto
+TextEncoderEncodeIntoResult TextEncoder::encode_into(String const& source, JS::Handle<JS::Object> const& destination) const
+{
+    auto& destination_array = static_cast<JS::Uint8Array&>(*destination);
+    auto data = destination_array.data();
+
+    // 1. Let read be 0.
+    unsigned long long read = 0;
+    // 2. Let written be 0.
+    unsigned long long written = 0;
+
+    // NOTE: The AK::String is always UTF-8, so most of these steps are no-ops.
+    // 3. Let encoder be an instance of the UTF-8 encoder.
+    // 4. Let unused be the I/O queue of scalar values « end-of-queue ».
+    // 5. Convert source to an I/O queue of scalar values.
+    auto code_points = source.code_points();
+    auto it = code_points.begin();
+
+    // 6. While true:
+    while (true) {
+        // 6.1. Let item be the result of reading from source.
+        // 6.2. Let result be the result of running encoder’s handler on unused and item.
+        // 6.3. If result is finished, then break.
+        if (it.done())
+            break;
+        auto item = *it;
+        auto result = it.underlying_code_point_bytes();
+
+        // 6.4. Otherwise:
+        // 6.4.1. If destination’s byte length − written is greater than or equal to the number of bytes in result, then:
+        if (data.size() - written >= result.size()) {
+            // 6.4.1.1. If item is greater than U+FFFF, then increment read by 2.
+            if (item > 0xffff) {
+                read += 2;
+            }
+            // 6.4.1.2. Otherwise, increment read by 1.
+            else {
+                read++;
+            }
+
+            // 6.4.1.3. Write the bytes in result into destination, with startingOffset set to written.
+            // 6.4.1.4. Increment written by the number of bytes in result.
+            for (auto byte : result)
+                data[written++] = byte;
+        }
+        // 6.4.2. Otherwise, break.
+        else {
+            break;
+        }
+
+        ++it;
+    }
+
+    // 7. Return «[ "read" → read, "written" → written ]».
+    return { read, written };
+}
+
 // https://encoding.spec.whatwg.org/#dom-textencoder-encoding
 FlyString const& TextEncoder::encoding()
 {
-    static const FlyString encoding = "utf-8"_fly_string;
+    static FlyString const encoding = "utf-8"_fly_string;
     return encoding;
 }
 

Userland/Libraries/LibWeb/Encoding/TextEncoder.h

@@ -15,6 +15,12 @@
 
 namespace Web::Encoding {
 
+// https://encoding.spec.whatwg.org/#dictdef-textencoderencodeintoresult
+struct TextEncoderEncodeIntoResult {
+    unsigned long long read;
+    unsigned long long written;
+};
+
 // https://encoding.spec.whatwg.org/#textencoder
 class TextEncoder final : public Bindings::PlatformObject {
     WEB_PLATFORM_OBJECT(TextEncoder, Bindings::PlatformObject);
@@ -24,7 +30,8 @@ public:
 
     virtual ~TextEncoder() override;
 
-    JS::Uint8Array* encode(String const& input) const;
+    JS::NonnullGCPtr<JS::Uint8Array> encode(String const& input) const;
+    TextEncoderEncodeIntoResult encode_into(String const& source, JS::Handle<JS::Object> const& destination) const;
 
     static FlyString const& encoding();
 

Userland/Libraries/LibWeb/Encoding/TextEncoder.idl

@@ -1,9 +1,20 @@
-[Exposed=(Window,Worker)]
+// https://encoding.spec.whatwg.org/#textencodercommon
+interface mixin TextEncoderCommon {
+    readonly attribute DOMString encoding;
+};
+
+// https://encoding.spec.whatwg.org/#dictdef-textencoderencodeintoresult
+dictionary TextEncoderEncodeIntoResult {
+    unsigned long long read;
+    unsigned long long written;
+};
+
+// https://encoding.spec.whatwg.org/#textencoder
+[Exposed=*]
 interface TextEncoder {
     constructor();
 
     [NewObject] Uint8Array encode(optional USVString input = "");
-    // TextEncoderEncodeIntoResult encodeInto(USVString source, [AllowShared] Uint8Array destination);
-
-    readonly attribute DOMString encoding;
+    TextEncoderEncodeIntoResult encodeInto(USVString source, [AllowShared] Uint8Array destination);
 };
+TextEncoder includes TextEncoderCommon;

Userland/Libraries/LibWeb/Forward.h

@@ -258,6 +258,7 @@ struct TextDecodeOptions;
 class TextDecoder;
 struct TextDecoderOptions;
 class TextEncoder;
+struct TextEncoderEncodeIntoResult;
 }
 
 namespace Web::Fetch {