LibWeb: Implement and use "isomorphic decoding"

Userland/Libraries/LibWeb/DOM/Document.cpp

@@ -347,7 +347,7 @@ WebIDL::ExceptionOr<JS::NonnullGCPtr<Document>> Document::create_and_initialize(
     // 16. If navigationParams's response has a `Refresh` header, then:
     if (auto maybe_refresh = navigation_params.response->header_list()->get("Refresh"sv.bytes()); maybe_refresh.has_value()) {
         // 1. Let value be the isomorphic decoding of the value of the header.
-        auto const& value = maybe_refresh.value();
+        auto value = Infra::isomorphic_decode(maybe_refresh.value());
 
         // 2. Run the shared declarative refresh steps with document and value.
         document->shared_declarative_refresh_steps(value, nullptr);

Userland/Libraries/LibWeb/Fetch/Infrastructure/HTTP/Headers.cpp

@@ -20,6 +20,7 @@
 #include <LibWeb/Fetch/Infrastructure/HTTP/Headers.h>
 #include <LibWeb/Fetch/Infrastructure/HTTP/Methods.h>
 #include <LibWeb/Infra/ByteSequences.h>
+#include <LibWeb/Infra/Strings.h>
 #include <LibWeb/Loader/ResourceLoader.h>
 #include <LibWeb/MimeSniff/MimeType.h>
 
@@ -52,8 +53,8 @@ Header Header::copy(Header const& header)
 Header Header::from_string_pair(StringView name, StringView value)
 {
     return Header {
-        .name = MUST(ByteBuffer::copy(name.bytes())),
-        .value = MUST(ByteBuffer::copy(value.bytes())),
+        .name = Infra::isomorphic_encode(name),
+        .value = Infra::isomorphic_encode(value),
     };
 }
 
@@ -135,7 +136,7 @@ Optional<Vector<String>> get_decode_and_split_header_value(ReadonlyBytes value)
     // To get, decode, and split a header value value, run these steps:
 
     // 1. Let input be the result of isomorphic decoding value.
-    auto input = StringView { value };
+    auto input = Infra::isomorphic_decode(value);
 
     // 2. Let position be a position variable for input, initially pointing at the start of input.
     auto lexer = GenericLexer { input };
@@ -530,7 +531,8 @@ bool is_cors_safelisted_request_header(Header const& header)
             return false;
 
         // 2. Let mimeType be the result of parsing the result of isomorphic decoding value.
-        auto mime_type = MimeSniff::MimeType::parse(StringView { value });
+        auto decoded = Infra::isomorphic_decode(value);
+        auto mime_type = MimeSniff::MimeType::parse(decoded);
 
         // 3. If mimeType is failure, then return false.
         if (!mime_type.has_value())
@@ -733,6 +735,7 @@ bool is_forbidden_request_header(Header const& header)
         auto parsed_values = get_decode_and_split_header_value(header.value);
 
         // 2. For each method of parsedValues: if the isomorphic encoding of method is a forbidden method, then return true.
+        // Note: The values returned from get_decode_and_split_header_value have already been decoded.
         if (parsed_values.has_value() && any_of(*parsed_values, [](auto method) { return is_forbidden_method(method.bytes()); }))
             return true;
     }
@@ -833,10 +836,10 @@ Variant<Vector<ByteBuffer>, ExtractHeaderParseFailure, Empty> extract_header_lis
 Optional<RangeHeaderValue> parse_single_range_header_value(ReadonlyBytes value)
 {
     // 1. Let data be the isomorphic decoding of value.
-    auto data = StringView { value };
+    auto data = Infra::isomorphic_decode(value);
 
     // 2. If data does not start with "bytes=", then return failure.
-    if (!data.starts_with("bytes="sv))
+    if (!data.starts_with_bytes("bytes="sv))
         return {};
 
     // 3. Let position be a position variable for data, initially pointing at the 6th code point of data.

Userland/Libraries/LibWeb/Fetch/Infrastructure/URL.cpp

@@ -9,6 +9,7 @@
 #include <AK/Base64.h>
 #include <LibURL/URL.h>
 #include <LibWeb/Fetch/Infrastructure/URL.h>
+#include <LibWeb/Infra/Strings.h>
 #include <LibWeb/MimeSniff/MimeType.h>
 
 namespace Web::Fetch::Infrastructure {
@@ -75,7 +76,7 @@ ErrorOr<DataURL> process_data_url(URL::URL const& data_url)
         trimmed_substring_view = trimmed_substring_view.trim(" "sv, TrimMode::Right);
         if (trimmed_substring_view.ends_with(';')) {
             // 1. Let stringBody be the isomorphic decode of body.
-            auto string_body = StringView(body);
+            auto string_body = Infra::isomorphic_decode(body);
 
             // 2. Set body to the forgiving-base64 decode of stringBody.
             // 3. If body is failure, then return failure.

Userland/Libraries/LibWeb/HTML/WindowOrWorkerGlobalScope.cpp

@@ -13,7 +13,6 @@
 #include <AK/Vector.h>
 #include <LibJS/Heap/HeapFunction.h>
 #include <LibJS/Runtime/Array.h>
-#include <LibTextCodec/Decoder.h>
 #include <LibWeb/Bindings/MainThreadVM.h>
 #include <LibWeb/Crypto/Crypto.h>
 #include <LibWeb/Fetch/FetchMethod.h>
@@ -36,6 +35,7 @@
 #include <LibWeb/HighResolutionTime/Performance.h>
 #include <LibWeb/HighResolutionTime/SupportedPerformanceTypes.h>
 #include <LibWeb/IndexedDB/IDBFactory.h>
+#include <LibWeb/Infra/Strings.h>
 #include <LibWeb/PerformanceTimeline/EntryTypes.h>
 #include <LibWeb/PerformanceTimeline/PerformanceObserver.h>
 #include <LibWeb/PerformanceTimeline/PerformanceObserverEntryList.h>
@@ -142,10 +142,8 @@ WebIDL::ExceptionOr<String> WindowOrWorkerGlobalScopeMixin::atob(String const& d
         return WebIDL::InvalidCharacterError::create(realm, "Input string is not valid base64 data"_string);
 
     // 3. Return decodedData.
-    // decode_base64() returns a byte buffer. LibJS uses UTF-8 for strings. Use Latin1Decoder to convert bytes 128-255 to UTF-8.
-    auto decoder = TextCodec::decoder_for_exact_name("ISO-8859-1"sv);
-    VERIFY(decoder.has_value());
-    return TRY_OR_THROW_OOM(vm, decoder->to_utf8(decoded_data.value()));
+    // decode_base64() returns a byte buffer. LibJS uses UTF-8 for strings. Use isomorphic decoding to convert bytes to UTF-8.
+    return Infra::isomorphic_decode(decoded_data.value());
 }
 
 // https://html.spec.whatwg.org/multipage/timers-and-user-prompts.html#dom-queuemicrotask

Userland/Libraries/LibWeb/Infra/Strings.cpp

@@ -144,4 +144,27 @@ ErrorOr<String> to_ascii_uppercase(StringView string)
     return string_builder.to_string();
 }
 
+// https://infra.spec.whatwg.org/#isomorphic-encode
+ByteBuffer isomorphic_encode(StringView input)
+{
+    ByteBuffer buf = {};
+    for (auto code_point : Utf8View { input }) {
+        // VERIFY(code_point <= 0xFF);
+        if (code_point > 0xFF)
+            dbgln("FIXME: Trying to isomorphic encode a string with code points > U+00FF.");
+        buf.append((u8)code_point);
+    }
+    return buf;
+}
+
+// https://infra.spec.whatwg.org/#isomorphic-decode
+String isomorphic_decode(ReadonlyBytes input)
+{
+    StringBuilder builder(input.size());
+    for (u8 code_point : input) {
+        builder.append_code_point(code_point);
+    }
+    return builder.to_string_without_validation();
+}
+
 }

Userland/Libraries/LibWeb/Infra/Strings.h

@@ -20,5 +20,7 @@ bool is_code_unit_prefix(StringView potential_prefix, StringView input);
 ErrorOr<String> convert_to_scalar_value_string(StringView string);
 ErrorOr<String> to_ascii_lowercase(StringView string);
 ErrorOr<String> to_ascii_uppercase(StringView string);
+ByteBuffer isomorphic_encode(StringView input);
+String isomorphic_decode(ReadonlyBytes input);
 
 }