mirrors/ladybird
0
0
Fork 0
mirror of https://github.com/LadybirdBrowser/ladybird.git synced 2025-05-10 21:22:53 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 9

AK+LibJS: Implement String.from{CharCode,CodePoint} using UTF-16 strings

Browse source 
Most of String.prototype and RegExp.prototype is implemented with UTF-16
so this is to prevent extra copying of the string data.
This commit is contained in:
Timothy Flynn 2021-08-02 17:02:17 -04:00 • committed by Andreas Kling
parent b6ff7f4fcc
commit 70080feab2
Notes: sideshowbarker 2024-07-18 07:30:31 +09:00
3 changed files with 31 additions and 18 deletions
Download patch file Download diff file Expand all files Collapse all files

24
AK/Utf16View.cpp
View file

@ -25,15 +25,8 @@ static Vector<u16> to_utf16_impl(UtfViewType const& view) requires(IsSame<UtfVie
{
Vector<u16> utf16_data;
for (auto code_point : view) {
if (code_point < first_supplementary_plane_code_point) {
utf16_data.append(static_cast<u16>(code_point));
} else {
code_point -= first_supplementary_plane_code_point;
utf16_data.append(static_cast<u16>(high_surrogate_min | (code_point >> 10)));
utf16_data.append(static_cast<u16>(low_surrogate_min | (code_point & 0x3ff)));
}
}
for (auto code_point : view)
code_point_to_utf16(utf16_data, code_point);
return utf16_data;
}
@ -53,6 +46,19 @@ Vector<u16> utf32_to_utf16(Utf32View const& utf32_view)
return to_utf16_impl(utf32_view);
}
void code_point_to_utf16(Vector<u16>& string, u32 code_point)
{
VERIFY(is_unicode(code_point));
if (code_point < first_supplementary_plane_code_point) {
string.append(static_cast<u16>(code_point));
} else {
code_point -= first_supplementary_plane_code_point;
string.append(static_cast<u16>(high_surrogate_min | (code_point >> 10)));
string.append(static_cast<u16>(low_surrogate_min | (code_point & 0x3ff)));
}
}
bool Utf16View::is_high_surrogate(u16 code_unit)
{
return (code_unit >= high_surrogate_min) && (code_unit <= high_surrogate_max);