ef6753a9f9
Add explicit IgnoreBOM and ErrorMode options to LibTextCodec decoders, and thread them through TextDecoder and TextDecoderStream. This lets Web-facing decoder APIs preserve BOMs when requested and use fatal error handling without post-processing decoded output. NB: RemoveBOM was renamed to IgnoreBOM as "RemoveBOM" is the name used by encoding_rs and was previously an implementation detail. The new name matches what is used by the encoding standard as it is now also used in LibWeb.
87 lines
3.8 KiB
C++
87 lines
3.8 KiB
C++
/*
|
||
* Copyright (c) 2022, Ali Mohammad Pur <mpfard@serenityos.org>
|
||
* Copyright (c) 2024, Simon Wanner <simon@skyrising.xyz>
|
||
*
|
||
* SPDX-License-Identifier: BSD-2-Clause
|
||
*/
|
||
|
||
#include <AK/FlyString.h>
|
||
#include <LibJS/Runtime/TypedArray.h>
|
||
#include <LibWeb/Bindings/Intrinsics.h>
|
||
#include <LibWeb/Bindings/TextDecoder.h>
|
||
#include <LibWeb/Encoding/TextDecoder.h>
|
||
#include <LibWeb/WebIDL/AbstractOperations.h>
|
||
#include <LibWeb/WebIDL/Buffers.h>
|
||
|
||
namespace Web::Encoding {
|
||
|
||
GC_DEFINE_ALLOCATOR(TextDecoder);
|
||
|
||
// https://encoding.spec.whatwg.org/#dom-textdecoder
|
||
WebIDL::ExceptionOr<GC::Ref<TextDecoder>> TextDecoder::construct_impl(JS::Realm& realm, StringView label, Bindings::TextDecoderOptions const& options)
|
||
{
|
||
auto& vm = realm.vm();
|
||
|
||
// 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, TRY_OR_THROW_OOM(vm, String::formatted("Invalid encoding {}", label)) };
|
||
|
||
// 3. Set this’s encoding to encoding.
|
||
// https://encoding.spec.whatwg.org/#dom-textdecoder-encoding
|
||
// The encoding getter steps are to return this’s encoding’s name, ASCII lowercased.
|
||
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 ? TextCodec::ErrorMode::Fatal : TextCodec::ErrorMode::Replacement;
|
||
|
||
// 5. Set this’s ignore BOM to options["ignoreBOM"].
|
||
auto ignore_bom = options.ignore_bom;
|
||
|
||
// NOTE: This should happen in decode(), but we don't support streaming yet and share decoders across calls.
|
||
auto decoder = TextCodec::decoder_for_exact_name(encoding.value());
|
||
VERIFY(decoder.has_value());
|
||
|
||
return realm.create<TextDecoder>(realm, *decoder, lowercase_encoding_name, error_mode, ignore_bom);
|
||
}
|
||
|
||
// https://encoding.spec.whatwg.org/#dom-textdecoder
|
||
TextDecoder::TextDecoder(JS::Realm& realm, TextCodec::Decoder& decoder, FlyString encoding, TextCodec::ErrorMode error_mode, bool ignore_bom)
|
||
: PlatformObject(realm)
|
||
, TextDecoderCommonMixin(decoder, move(encoding), error_mode, ignore_bom)
|
||
{
|
||
}
|
||
|
||
TextDecoder::~TextDecoder() = default;
|
||
|
||
void TextDecoder::initialize(JS::Realm& realm)
|
||
{
|
||
WEB_SET_PROTOTYPE_FOR_INTERFACE(TextDecoder);
|
||
Base::initialize(realm);
|
||
}
|
||
|
||
// https://encoding.spec.whatwg.org/#dom-textdecoder-decode
|
||
WebIDL::ExceptionOr<String> TextDecoder::decode(Optional<WebIDL::BufferSourceVariant> input, Optional<Bindings::TextDecodeOptions> const&) const
|
||
{
|
||
auto ignore_bom = m_ignore_bom ? TextCodec::IgnoreBOM::Yes : TextCodec::IgnoreBOM::No;
|
||
if (!input.has_value()) {
|
||
auto result = m_decoder.to_utf8({}, ignore_bom, m_error_mode);
|
||
if (result.is_error() && result.error().code() != ENOMEM)
|
||
return WebIDL::SimpleException { WebIDL::SimpleExceptionType::TypeError, "Decoding failed"sv };
|
||
return TRY_OR_THROW_OOM(vm(), move(result));
|
||
}
|
||
|
||
// FIXME: Implement the streaming stuff.
|
||
auto data_buffer_or_error = WebIDL::get_buffer_source_copy(*input);
|
||
if (data_buffer_or_error.is_error())
|
||
return WebIDL::OperationError::create(realm(), "Failed to copy bytes from ArrayBuffer"_utf16);
|
||
auto& data_buffer = data_buffer_or_error.value();
|
||
auto result = m_decoder.to_utf8({ data_buffer.data(), data_buffer.size() }, ignore_bom, m_error_mode);
|
||
if (result.is_error() && result.error().code() != ENOMEM)
|
||
return WebIDL::SimpleException { WebIDL::SimpleExceptionType::TypeError, "Decoding failed"sv };
|
||
return TRY_OR_THROW_OOM(vm(), move(result));
|
||
}
|
||
|
||
}
|