RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-05-31 23:38:12 +00:00
Code Issues Activity Actions

LibJS: Decode UTF-16 surrogate pairs during string literal construction

Browse source 
Rather than deferring this decoding to PrimitiveString, we can decode
surrogate pairs when parsing the string. This prevents a string copy
when constructing the PrimitiveString.
This commit is contained in:
Timothy Flynn 2021-08-02 10:06:42 -04:00 • committed by Andreas Kling
parent 27d555bab0
commit 4c2cc419f9
2 changed files with 28 additions and 14 deletions
Download patch file Download diff file Expand all files Collapse all files

11
Userland/Libraries/LibJS/Runtime/PrimitiveString.cpp
View file

@ -63,17 +63,6 @@ PrimitiveString* js_string(Heap& heap, String string)
return &heap.vm().single_ascii_character_string(ch);
}
// UTF-8 strings must first be transcoded to UTF-16, even though they are stored as String objects
// internally, to parse encoded surrogate pairs. As an optimization to reduce string copying, only
// perform that transcoding if there are non-ASCII codepoints in the string.
for (auto it : string) {
auto ch = static_cast<u8>(it);
if (!is_ascii(ch)) {
auto utf16_string = AK::utf8_to_utf16(string);
return js_string(heap, Utf16View { utf16_string });
}
}
return heap.allocate_without_global_object<PrimitiveString>(move(string));
}