LibJS+LibUnicode: Port Intl.ListFormat to UTF-16 strings

Author: https://github.com/trflynn89
Commit: db2148b44a
Pull-request: https://github.com/LadybirdBrowser/ladybird/pull/5581

Libraries/LibJS/Runtime/Intl/DurationFormat.cpp

@@ -731,13 +731,13 @@ Vector<DurationFormatPart> list_format_parts(VM& vm, DurationFormat const& durat
     auto list_format = construct_list_format(vm, duration_format, list_format_options);
 
     // 7. Let strings be a new empty List.
-    Vector<String> strings;
+    Vector<Utf16String> strings;
     strings.ensure_capacity(partitioned_parts_list.size());
 
     // 8. For each element parts of partitionedPartsList, do
     for (auto const& parts : partitioned_parts_list) {
         // a. Let string be the empty String.
-        StringBuilder string;
+        StringBuilder string(StringBuilder::Mode::UTF16);
 
         // b. For each Record { [[Type]], [[Value]], [[Unit]] } part in parts, do
         for (auto const& part : parts) {
@@ -746,7 +746,7 @@ Vector<DurationFormatPart> list_format_parts(VM& vm, DurationFormat const& durat
         }
 
         // c. Append string to strings.
-        strings.unchecked_append(MUST(string.to_string()));
+        strings.unchecked_append(string.to_utf16_string());
     }
 
     // 9. Let formattedPartsList be CreatePartsFromList(lf, strings).
@@ -786,7 +786,7 @@ Vector<DurationFormatPart> list_format_parts(VM& vm, DurationFormat const& durat
             VERIFY(list_part.type == "literal"sv);
 
             // ii. Append the Record { [[Type]]: "literal", [[Value]]: listPart.[[Value]], [[Unit]]: empty } to flattenedPartsList.
-            flattened_parts_list.append({ .type = "literal"sv, .value = move(list_part.value), .unit = {} });
+            flattened_parts_list.append({ .type = "literal"sv, .value = list_part.value.to_utf8_but_should_be_ported_to_utf16(), .unit = {} });
         }
     }
 

Libraries/LibJS/Runtime/Intl/ListFormat.cpp

@@ -35,13 +35,13 @@ ReadonlySpan<ResolutionOptionDescriptor> ListFormat::resolution_option_descripto
 }
 
 // 14.5.2 CreatePartsFromList ( listFormat, list ), https://tc39.es/ecma402/#sec-createpartsfromlist
-Vector<Unicode::ListFormat::Partition> create_parts_from_list(ListFormat const& list_format, ReadonlySpan<String> list)
+Vector<Unicode::ListFormat::Partition> create_parts_from_list(ListFormat const& list_format, ReadonlySpan<Utf16String> list)
 {
     return list_format.formatter().format_to_parts(list);
 }
 
 // 14.5.3 FormatList ( listFormat, list ), https://tc39.es/ecma402/#sec-formatlist
-String format_list(ListFormat const& list_format, ReadonlySpan<String> list)
+Utf16String format_list(ListFormat const& list_format, ReadonlySpan<Utf16String> list)
 {
     // 1. Let parts be ! CreatePartsFromList(listFormat, list).
     // 2. Let result be the empty String.
@@ -52,7 +52,7 @@ String format_list(ListFormat const& list_format, ReadonlySpan<String> list)
 }
 
 // 14.5.4 FormatListToParts ( listFormat, list ), https://tc39.es/ecma402/#sec-formatlisttoparts
-GC::Ref<Array> format_list_to_parts(VM& vm, ListFormat const& list_format, ReadonlySpan<String> list)
+GC::Ref<Array> format_list_to_parts(VM& vm, ListFormat const& list_format, ReadonlySpan<Utf16String> list)
 {
     auto& realm = *vm.current_realm();
 
@@ -88,19 +88,19 @@ GC::Ref<Array> format_list_to_parts(VM& vm, ListFormat const& list_format, Reado
 }
 
 // 14.5.5 StringListFromIterable ( iterable ), https://tc39.es/ecma402/#sec-createstringlistfromiterable
-ThrowCompletionOr<Vector<String>> string_list_from_iterable(VM& vm, Value iterable)
+ThrowCompletionOr<Vector<Utf16String>> string_list_from_iterable(VM& vm, Value iterable)
 {
     // 1. If iterable is undefined, then
     if (iterable.is_undefined()) {
         // a. Return a new empty List.
-        return Vector<String> {};
+        return Vector<Utf16String> {};
     }
 
     // 2. Let iteratorRecord be ? GetIterator(iterable, sync).
     auto iterator_record = TRY(get_iterator(vm, iterable, IteratorHint::Sync));
 
     // 3. Let list be a new empty List.
-    Vector<String> list;
+    Vector<Utf16String> list;
 
     // 4. Repeat,
     while (true) {
@@ -123,7 +123,7 @@ ThrowCompletionOr<Vector<String>> string_list_from_iterable(VM& vm, Value iterab
         }
 
         // iii. Append next to list.
-        list.append(next->as_string().utf8_string());
+        list.append(next->as_string().utf16_string());
     }
 }
 

Libraries/LibJS/Runtime/Intl/ListFormat.h

@@ -57,9 +57,9 @@ private:
     OwnPtr<Unicode::ListFormat> m_formatter;
 };
 
-Vector<Unicode::ListFormat::Partition> create_parts_from_list(ListFormat const&, ReadonlySpan<String> list);
+Vector<Unicode::ListFormat::Partition> create_parts_from_list(ListFormat const&, ReadonlySpan<Utf16String> list);
-String format_list(ListFormat const&, ReadonlySpan<String> list);
+Utf16String format_list(ListFormat const&, ReadonlySpan<Utf16String> list);
-GC::Ref<Array> format_list_to_parts(VM&, ListFormat const&, ReadonlySpan<String> list);
+GC::Ref<Array> format_list_to_parts(VM&, ListFormat const&, ReadonlySpan<Utf16String> list);
-ThrowCompletionOr<Vector<String>> string_list_from_iterable(VM&, Value iterable);
+ThrowCompletionOr<Vector<Utf16String>> string_list_from_iterable(VM&, Value iterable);
 
 }

Libraries/LibUnicode/ICU.cpp

@@ -138,16 +138,13 @@ TimeZoneData::TimeZoneData(NonnullOwnPtr<icu::TimeZone> time_zone)
 {
 }
 
-Vector<icu::UnicodeString> icu_string_list(ReadonlySpan<String> strings)
+Vector<icu::UnicodeString> icu_string_list(ReadonlySpan<Utf16String> strings)
 {
     Vector<icu::UnicodeString> result;
     result.ensure_capacity(strings.size());
 
-    for (auto const& string : strings) {
+    for (auto const& string : strings)
-        auto view = string.bytes_as_string_view();
+        result.unchecked_append(icu_string(string));
-        icu::UnicodeString icu_string(view.characters_without_null_termination(), static_cast<i32>(view.length()));
-        result.unchecked_append(move(icu_string));
-    }
 
     return result;
 }

Libraries/LibUnicode/ICU.h

@@ -99,7 +99,16 @@ ALWAYS_INLINE icu::UnicodeString icu_string(StringView string)
     return icu::UnicodeString::fromUTF8(icu_string_piece(string));
 }
 
-Vector<icu::UnicodeString> icu_string_list(ReadonlySpan<String> strings);
+// If the Utf16View has ASCII storage, this creates an owned icu::UnicodeString. Otherwise, the icu::UnicodeString is
+// unowned (i.e. it is effectively a view).
+ALWAYS_INLINE icu::UnicodeString icu_string(Utf16View const& string)
+{
+    if (string.has_ascii_storage())
+        return icu::UnicodeString::fromUTF8(icu_string_piece(string.bytes()));
+    return { false, string.utf16_span().data(), static_cast<i32>(string.length_in_code_units()) };
+}
+
+Vector<icu::UnicodeString> icu_string_list(ReadonlySpan<Utf16String> strings);
 
 String icu_string_to_string(icu::UnicodeString const& string);
 String icu_string_to_string(UChar const*, i32 length);

Libraries/LibUnicode/ListFormat.cpp

@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2024, Tim Flynn <trflynn89@serenityos.org>
+ * Copyright (c) 2024-2025, Tim Flynn <trflynn89@ladybird.org>
  *
  * SPDX-License-Identifier: BSD-2-Clause
  */
@@ -82,7 +82,7 @@ public:
 
     virtual ~ListFormatImpl() override = default;
 
-    virtual String format(ReadonlySpan<String> list) const override
+    virtual Utf16String format(ReadonlySpan<Utf16String> list) const override
     {
         UErrorCode status = U_ZERO_ERROR;
 
@@ -94,10 +94,10 @@ public:
         if (icu_failure(status))
             return {};
 
-        return icu_string_to_string(formatted_string);
+        return icu_string_to_utf16_string(formatted_string);
     }
 
-    virtual Vector<Partition> format_to_parts(ReadonlySpan<String> list) const override
+    virtual Vector<Partition> format_to_parts(ReadonlySpan<Utf16String> list) const override
     {
         UErrorCode status = U_ZERO_ERROR;
 
@@ -118,14 +118,14 @@ public:
             auto type = icu_list_format_field_to_string(position.getField());
             auto part = formatted_string.tempSubStringBetween(position.getStart(), position.getLimit());
 
-            result.empend(type, icu_string_to_string(part));
+            result.empend(type, icu_string_to_utf16_string(part));
         }
 
         return result;
     }
 
 private:
-    Optional<icu::FormattedList> format_list_impl(ReadonlySpan<String> list) const
+    Optional<icu::FormattedList> format_list_impl(ReadonlySpan<Utf16String> list) const
     {
         UErrorCode status = U_ZERO_ERROR;
 

Libraries/LibUnicode/ListFormat.h

@@ -1,12 +1,12 @@
 /*
- * Copyright (c) 2024, Tim Flynn <trflynn89@serenityos.org>
+ * Copyright (c) 2024-2025, Tim Flynn <trflynn89@ladybird.org>
  *
  * SPDX-License-Identifier: BSD-2-Clause
  */
 
 #pragma once
 
-#include <AK/String.h>
+#include <AK/Utf16String.h>
 #include <AK/Vector.h>
 #include <LibUnicode/Locale.h>
 
@@ -27,11 +27,11 @@ public:
 
     struct Partition {
         StringView type;
-        String value;
+        Utf16String value;
     };
 
-    virtual String format(ReadonlySpan<String> list) const = 0;
+    virtual Utf16String format(ReadonlySpan<Utf16String> list) const = 0;
-    virtual Vector<Partition> format_to_parts(ReadonlySpan<String> list) const = 0;
+    virtual Vector<Partition> format_to_parts(ReadonlySpan<Utf16String> list) const = 0;
 
 protected:
     ListFormat() = default;