/* * Copyright (c) 2026-present, the Ladybird developers. * * SPDX-License-Identifier: BSD-2-Clause */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include namespace Web::Encoding { GC_DEFINE_ALLOCATOR(TextDecoderStream); // https://encoding.spec.whatwg.org/#dom-textdecoderstream WebIDL::ExceptionOr> TextDecoderStream::construct_impl(JS::Realm& realm, FlyString label, Bindings::TextDecoderOptions const& options) { // 1. Let encoding be the result of getting an encoding from label. auto encoding = TextCodec::get_standardized_encoding(label); // 2. If encoding is failure or replacement, then throw a RangeError. if (!encoding.has_value() || encoding->equals_ignoring_ascii_case("replacement"sv)) return WebIDL::SimpleException { WebIDL::SimpleExceptionType::RangeError, MUST(String::formatted("Invalid encoding {}", label)) }; // 3. Set this’s encoding to encoding. auto lowercase_encoding_name = encoding.value().to_ascii_lowercase_string(); // 4. If options["fatal"] is true, then set this’s error mode to "fatal". auto error_mode = options.fatal ? ErrorMode::Fatal : ErrorMode::Replacement; // 5. Set this’s ignore BOM to options["ignoreBOM"]. auto ignore_bom = options.ignore_bom; // 6. Set this’s decoder to a new instance of this’s encoding’s decoder, and set this’s I/O queue to a new I/O queue. auto decoder = TextCodec::decoder_for_exact_name(encoding.value()); VERIFY(decoder.has_value()); // NB: Steps 7-11 — we create the TransformStream and the TextDecoderStream first so that we can refer to the // stream from the transform/flush algorithms. // 9. Let transformStream be a new TransformStream. auto transform_stream = realm.create(realm); auto stream = realm.create(realm, transform_stream, *decoder, lowercase_encoding_name, error_mode, ignore_bom); // 7. Let transformAlgorithm be an algorithm which takes a chunk argument and runs the decode and enqueue a chunk // algorithm with this and chunk. auto transform_algorithm = GC::create_function(realm.heap(), [stream](JS::Value chunk) -> GC::Ref { auto& realm = stream->realm(); if (auto result = stream->decode_and_enqueue_chunk(chunk); result.is_error()) return WebIDL::create_rejected_promise_from_exception(realm, result.release_error()); return WebIDL::create_resolved_promise(realm, JS::js_undefined()); }); // 8. Let flushAlgorithm be an algorithm which takes no arguments and runs the flush and enqueue algorithm with this. auto flush_algorithm = GC::create_function(realm.heap(), [stream]() -> GC::Ref { auto& realm = stream->realm(); if (auto result = stream->flush_and_enqueue(); result.is_error()) return WebIDL::create_rejected_promise_from_exception(realm, result.release_error()); return WebIDL::create_resolved_promise(realm, JS::js_undefined()); }); // 10. Set up transformStream with transformAlgorithm set to transformAlgorithm and flushAlgorithm set to flushAlgorithm. transform_stream->set_up(transform_algorithm, flush_algorithm); // 11. Set this’s transform to transformStream. // NB: Done via the GenericTransformStreamMixin constructor above. return stream; } TextDecoderStream::TextDecoderStream(JS::Realm& realm, GC::Ref transform, TextCodec::Decoder& decoder, FlyString encoding, ErrorMode error_mode, bool ignore_bom) : Bindings::PlatformObject(realm) , Streams::GenericTransformStreamMixin(transform) , TextDecoderCommonMixin(decoder, move(encoding), error_mode, ignore_bom) , m_streaming_decoder(make(m_encoding)) { } TextDecoderStream::~TextDecoderStream() = default; void TextDecoderStream::initialize(JS::Realm& realm) { WEB_SET_PROTOTYPE_FOR_INTERFACE(TextDecoderStream); Base::initialize(realm); } void TextDecoderStream::visit_edges(Cell::Visitor& visitor) { Base::visit_edges(visitor); Streams::GenericTransformStreamMixin::visit_edges(visitor); } // https://encoding.spec.whatwg.org/#decode-and-enqueue-a-chunk WebIDL::ExceptionOr TextDecoderStream::decode_and_enqueue_chunk(JS::Value chunk) { auto& realm = this->realm(); auto& vm = realm.vm(); // 1. Let bufferSource be the result of converting chunk to an AllowSharedBufferSource. if (!WebIDL::is_buffer_source_type(chunk)) return WebIDL::SimpleException { WebIDL::SimpleExceptionType::TypeError, "Chunk is not a BufferSource"sv }; // 2. Push a copy of bufferSource to decoder's I/O queue. auto buffer_or_error = WebIDL::get_buffer_source_copy(chunk.as_object()); if (buffer_or_error.is_error()) return WebIDL::OperationError::create(realm, "Failed to copy bytes from BufferSource"_utf16); auto buffer = buffer_or_error.release_value(); auto decoded = TRY_OR_THROW_OOM(vm, m_streaming_decoder->to_utf8(buffer.bytes())); // 3-4. Run "processing an item" until the input is exhausted, accumulating the output, then enqueue any non-empty // result. If processing returns error, throw a TypeError. return enqueue_decoded_output(decoded); } // https://encoding.spec.whatwg.org/#flush-and-enqueue WebIDL::ExceptionOr TextDecoderStream::flush_and_enqueue() { // 1-3. Drain decoder's I/O queue and run "processing an item" to completion. auto decoded = TRY_OR_THROW_OOM(vm(), m_streaming_decoder->finish()); return enqueue_decoded_output(decoded); } WebIDL::ExceptionOr TextDecoderStream::enqueue_decoded_output(String const& decoded) { auto& realm = this->realm(); auto& vm = realm.vm(); // https://encoding.spec.whatwg.org/#concept-td-serialize // FIXME: The underlying TextCodec decoders currently strip leading BOMs unconditionally for UTF-8 and UTF-16BE/LE, // so the "ignore BOM" flag is effectively ignored here. Once the decoders accept a "preserve BOM" mode, // plumb m_ignore_bom through and strip the BOM from `decoded` only when m_ignore_bom is false. if (!m_bom_seen && !decoded.is_empty()) m_bom_seen = true; if (decoded.is_empty()) return {}; // If decoder's error mode is "fatal" and processing produced any error, throw a TypeError. // NB: We can only detect this approximately by looking for U+FFFD in the decoded output, which the underlying // decoder substitutes for invalid sequences. This matches the existing TextDecoder.decode() behavior. if (fatal() && decoded.contains(0xFFFD)) return WebIDL::SimpleException { WebIDL::SimpleExceptionType::TypeError, "Decoding failed"sv }; auto js_string = JS::PrimitiveString::create(vm, Utf16String::from_utf8(decoded)); return Streams::transform_stream_default_controller_enqueue(*m_transform->controller(), js_string); } }