Incorrect length for some emojis

[sallar]

Hi there,
the library returns incorrect length:

len("❤️"); // 2
len("👊🏽"); // 2

I don’t seem to be able to find a solution for this in ANY library except lodash which has a complicated regex...

[sallar]

I fixed this problem in my own library using a regex borrowed from lodash...
https://github.com/sallar/stringz

[revelt]

If stringz issue sallar/stringz#27 does come through, we could replace aging astral-regex with that regex and results would be correct again in all the cases.

[Richienb]

Here's Lodash's strategy: https://github.com/lodash/lodash/blob/master/.internal/unicodeSize.js

[revelt]

In general, string length is string.length. If we start to calculate the code point count and yet call that a string length — that's a mistake. This package should have been called code point count to match the promise in the readme.

@sindresorhus If I may advise, if I was a maintainer of this program, I'd propose to can this project and leave emoji length calculation to specialist packages with proper support where people seriously spend time on Unicode and are able update the program to cater new-fangled emoji that is released continuously. This is not a quick one-liner package!

[Richienb]

string-length was made to count what the logically counted length of a string is. I've extracted part of Lodash's strategy which is considerably more maintained into another module. I'll open a PR for version 4.

[revelt]

I understand, yet terminology conundrum is still here, it's not string length, it's code point count