tc39
ES2019 summary shouldn't mention well-formed UTF-8 for JSON.stringify
In the spec intro:
Other updates included requiring that Array.prototype.sort be a stable sort, requiring that JSON.stringify return well-formed UTF-8 regardless of input, and clarifying Function.prototype.toString by requiring that it either return the corresponding original source text or a standard placeholder.
Strings are UTF-16 (https://tc39.es/ecma262/#sec-ecmascript-language-types-string-type), so a string-returning function cannot return UTF-8.
Maybe we should instead say "well-formed UTF-16", like we do for https://tc39.es/ecma262/#sec-isstringwellformedunicode, which I assume means "no unpaired surrogates"?
I'd prefer to describe the actual change, I think: "changing JSON.stringify to escape unpaired surrogates" or something like that. (Which, yes, was the actual change.) If we want to also explain the point of the change, I'd append "so that the resulting strings are always well-formed UTF-16".
