You can clone with
Seems like the \uXXXX codes are designed for the benefit of human authors.
We'll generate smaller files if we write UTF-8 directly.
well, after a half hour trying to figure out what we need to change, I realized that's harder to understand the source then I thought :(
I'll fix it. The change needs to happen in strenc, which is
responsible for encoding strings.
generate multi-byte UTF-8; closes #4
removing this test since it's fixed by kr/okjson#4