LibJS+AK: Use Vector<u16, 1> for UTF-16 string storage

It's very common to encounter single-character strings in JavaScript on
the web. We can make such strings significantly lighter by having a
1-character inline capacity on the Vectors.

7 files changed, 27 insertions, 26 deletions

diff --git a/AK/Utf16View.cpp b/AK/Utf16View.cpp
index d49ba18692..fcac008c01 100644
--- a/AK/Utf16View.cpp
+++ b/AK/Utf16View.cpp
@@ -21,9 +21,10 @@ static constexpr u32 replacement_code_point = 0xfffd;
 static constexpr u32 first_supplementary_plane_code_point = 0x10000;
 
 template<typename UtfViewType>
-static Vector<u16> to_utf16_impl(UtfViewType const& view) requires(IsSame<UtfViewType, Utf8View> || IsSame<UtfViewType, Utf32View>)
+static Vector<u16, 1> to_utf16_impl(UtfViewType const& view) requires(IsSame<UtfViewType, Utf8View> || IsSame<UtfViewType, Utf32View>)
 {
-    Vector<u16> utf16_data;
+    Vector<u16, 1> utf16_data;
+    utf16_data.ensure_capacity(view.length());
 
     for (auto code_point : view)
         code_point_to_utf16(utf16_data, code_point);
@@ -31,22 +32,22 @@ static Vector<u16> to_utf16_impl(UtfViewType const& view) requires(IsSame<UtfVie
     return utf16_data;
 }
 
-Vector<u16> utf8_to_utf16(StringView const& utf8_view)
+Vector<u16, 1> utf8_to_utf16(StringView const& utf8_view)
 {
     return to_utf16_impl(Utf8View { utf8_view });
 }
 
-Vector<u16> utf8_to_utf16(Utf8View const& utf8_view)
+Vector<u16, 1> utf8_to_utf16(Utf8View const& utf8_view)
 {
     return to_utf16_impl(utf8_view);
 }
 
-Vector<u16> utf32_to_utf16(Utf32View const& utf32_view)
+Vector<u16, 1> utf32_to_utf16(Utf32View const& utf32_view)
 {
     return to_utf16_impl(utf32_view);
 }
 
-void code_point_to_utf16(Vector<u16>& string, u32 code_point)
+void code_point_to_utf16(Vector<u16, 1>& string, u32 code_point)
 {
     VERIFY(is_unicode(code_point));
 
diff --git a/AK/Utf16View.h b/AK/Utf16View.h
index a332ed6a23..805ab08000 100644
--- a/AK/Utf16View.h
+++ b/AK/Utf16View.h
@@ -16,10 +16,10 @@
 
 namespace AK {
 
-Vector<u16> utf8_to_utf16(StringView const&);
-Vector<u16> utf8_to_utf16(Utf8View const&);
-Vector<u16> utf32_to_utf16(Utf32View const&);
-void code_point_to_utf16(Vector<u16>&, u32);
+Vector<u16, 1> utf8_to_utf16(StringView const&);
+Vector<u16, 1> utf8_to_utf16(Utf8View const&);
+Vector<u16, 1> utf32_to_utf16(Utf32View const&);
+void code_point_to_utf16(Vector<u16, 1>&, u32);
 
 class Utf16View;
 
diff --git a/Userland/Libraries/LibJS/Runtime/StringConstructor.cpp b/Userland/Libraries/LibJS/Runtime/StringConstructor.cpp
index cbac0b8d0d..d13a4f089f 100644
--- a/Userland/Libraries/LibJS/Runtime/StringConstructor.cpp
+++ b/Userland/Libraries/LibJS/Runtime/StringConstructor.cpp
@@ -140,7 +140,7 @@ JS_DEFINE_NATIVE_FUNCTION(StringConstructor::from_char_code)
 // 22.1.2.2 String.fromCodePoint ( ...codePoints ), https://tc39.es/ecma262/#sec-string.fromcodepoint
 JS_DEFINE_NATIVE_FUNCTION(StringConstructor::from_code_point)
 {
-    Vector<u16> string;
+    Vector<u16, 1> string;
     string.ensure_capacity(vm.argument_count()); // This will be an under-estimate if any code point is > 0xffff.
 
     for (size_t i = 0; i < vm.argument_count(); ++i) {
diff --git a/Userland/Libraries/LibJS/Runtime/Utf16String.cpp b/Userland/Libraries/LibJS/Runtime/Utf16String.cpp
index b2e40c252f..e5ad11e822 100644
--- a/Userland/Libraries/LibJS/Runtime/Utf16String.cpp
+++ b/Userland/Libraries/LibJS/Runtime/Utf16String.cpp
@@ -17,7 +17,7 @@ static NonnullRefPtr<Utf16StringImpl> the_empty_utf16_string()
     return empty_string;
 }
 
-Utf16StringImpl::Utf16StringImpl(Vector<u16> string)
+Utf16StringImpl::Utf16StringImpl(Vector<u16, 1> string)
     : m_string(move(string))
 {
 }
@@ -27,7 +27,7 @@ NonnullRefPtr<Utf16StringImpl> Utf16StringImpl::create()
     return adopt_ref(*new Utf16StringImpl());
 }
 
-NonnullRefPtr<Utf16StringImpl> Utf16StringImpl::create(Vector<u16> string)
+NonnullRefPtr<Utf16StringImpl> Utf16StringImpl::create(Vector<u16, 1> string)
 {
     return adopt_ref(*new Utf16StringImpl(move(string)));
 }
@@ -39,13 +39,13 @@ NonnullRefPtr<Utf16StringImpl> Utf16StringImpl::create(StringView const& string)
 
 NonnullRefPtr<Utf16StringImpl> Utf16StringImpl::create(Utf16View const& view)
 {
-    Vector<u16> string;
+    Vector<u16, 1> string;
     string.ensure_capacity(view.length_in_code_units());
     string.append(view.data(), view.length_in_code_units());
     return create(move(string));
 }
 
-Vector<u16> const& Utf16StringImpl::string() const
+Vector<u16, 1> const& Utf16StringImpl::string() const
 {
     return m_string;
 }
@@ -62,7 +62,7 @@ Utf16String::Utf16String()
 {
 }
 
-Utf16String::Utf16String(Vector<u16> string)
+Utf16String::Utf16String(Vector<u16, 1> string)
     : m_string(Detail::Utf16StringImpl::create(move(string)))
 {
 }
@@ -77,7 +77,7 @@ Utf16String::Utf16String(Utf16View const& string)
 {
 }
 
-Vector<u16> const& Utf16String::string() const
+Vector<u16, 1> const& Utf16String::string() const
 {
     return m_string->string();
 }
diff --git a/Userland/Libraries/LibJS/Runtime/Utf16String.h b/Userland/Libraries/LibJS/Runtime/Utf16String.h
index 640de51218..e38e5b4592 100644
--- a/Userland/Libraries/LibJS/Runtime/Utf16String.h
+++ b/Userland/Libraries/LibJS/Runtime/Utf16String.h
@@ -20,18 +20,18 @@ public:
     ~Utf16StringImpl() = default;
 
     static NonnullRefPtr<Utf16StringImpl> create();
-    static NonnullRefPtr<Utf16StringImpl> create(Vector<u16>);
+    static NonnullRefPtr<Utf16StringImpl> create(Vector<u16, 1>);
     static NonnullRefPtr<Utf16StringImpl> create(StringView const&);
     static NonnullRefPtr<Utf16StringImpl> create(Utf16View const&);
 
-    Vector<u16> const& string() const;
+    Vector<u16, 1> const& string() const;
     Utf16View view() const;
 
 private:
     Utf16StringImpl() = default;
-    explicit Utf16StringImpl(Vector<u16> string);
+    explicit Utf16StringImpl(Vector<u16, 1> string);
 
-    Vector<u16> m_string;
+    Vector<u16, 1> m_string;
 };
 
 }
@@ -39,11 +39,11 @@ private:
 class Utf16String {
 public:
     Utf16String();
-    explicit Utf16String(Vector<u16>);
+    explicit Utf16String(Vector<u16, 1>);
     explicit Utf16String(StringView const&);
     explicit Utf16String(Utf16View const&);
 
-    Vector<u16> const& string() const;
+    Vector<u16, 1> const& string() const;
     Utf16View view() const;
     Utf16View substring_view(size_t code_unit_offset, size_t code_unit_length) const;
     Utf16View substring_view(size_t code_unit_offset) const;
diff --git a/Userland/Libraries/LibRegex/RegexByteCode.cpp b/Userland/Libraries/LibRegex/RegexByteCode.cpp
index 231c076d10..2b6bc146c7 100644
--- a/Userland/Libraries/LibRegex/RegexByteCode.cpp
+++ b/Userland/Libraries/LibRegex/RegexByteCode.cpp
@@ -470,7 +470,7 @@ ALWAYS_INLINE ExecutionResult OpCode_Compare::execute(MatchInput const& input, M
                 return ExecutionResult::Failed_ExecuteLowPrioForks;
 
             Optional<String> str;
-            Vector<u16> utf16;
+            Vector<u16, 1> utf16;
             Vector<u32> data;
             data.ensure_capacity(length);
             for (size_t i = offset; i < offset + length; ++i)
@@ -557,7 +557,7 @@ ALWAYS_INLINE void OpCode_Compare::compare_char(MatchInput const& input, MatchSt
 
     auto input_view = input.view.substring_view(state.string_position, 1);
     Optional<String> str;
-    Vector<u16> utf16;
+    Vector<u16, 1> utf16;
     auto compare_view = input_view.construct_as_same({ &ch1, 1 }, str, utf16);
     bool equal;
     if (input.regex_options & AllFlags::Insensitive)
diff --git a/Userland/Libraries/LibRegex/RegexMatch.h b/Userland/Libraries/LibRegex/RegexMatch.h
index 60a1241eb3..6d707650ee 100644
--- a/Userland/Libraries/LibRegex/RegexMatch.h
+++ b/Userland/Libraries/LibRegex/RegexMatch.h
@@ -139,7 +139,7 @@ public:
         return view;
     }
 
-    RegexStringView construct_as_same(Span<u32> data, Optional<String>& optional_string_storage, Vector<u16>& optional_utf16_storage) const
+    RegexStringView construct_as_same(Span<u32> data, Optional<String>& optional_string_storage, Vector<u16, 1>& optional_utf16_storage) const
     {
         auto view = m_view.visit(
             [&]<typename T>(T const&) {