Docs of `charCodeAt` and `codePointAt` are flipped

Bug Report

Docs of charCodeAt and codePointAt are flipped. Described in MDN, the codePointAt returns Unicode code point, and charCodeAt returns UTF16 char code, however in JSDoc string of Typescript they are flipped.

How it is

https://github.com/microsoft/TypeScript/blob/89d05f7131b233c4482aaba5c15ad289f26173b4/lib/lib.es2015.core.d.ts#L401-L408

https://github.com/microsoft/TypeScript/blob/89d05f7131b233c4482aaba5c15ad289f26173b4/lib/lib.es5.d.ts#L404-L408

How it should be

The codePointAt() method returns a non-negative integer that is the Unicode code point value at the given position. Note that this function does not give the nth code point in a string, but the code point starting at the specified string index. Docs

The charCodeAt() method returns an integer between 0 and 65535 representing the UTF-16 code unit at the given index. Docs

[Ciantic]

That’s all quiet interesting, I will study that implementation.

Maybe the docs could mention at least that for all unicode points there exists a unicode character usually with a glyph. But for char codes that is not the case, as they are just bits.

[fatcerberus]

I’m pretty sure that will 1) break apart surrogates, and 2) cause more GC pressure creating temp arrays, unless the engine is specifically optimized for that case to avoid creating the temp.

Text processing is tricky.