- Migrate to AVA

- ZWJ support

Author: dotcypress

.eslintrc

@@ -1,7 +1,4 @@
 {
   "extends": ["standard"],
-  "plugins": ["mocha"],
-  "env": {
-    "mocha": true
-  }
+  "plugins": ["ava"]
 }

index.js

@@ -1,30 +1,37 @@
 'use strict'
 
-const HIGH_SURROGATE_START = 0xD800
-const HIGH_SURROGATE_END = 0xDBFF
+const HIGH_SURROGATE_START = 0xd800
+const HIGH_SURROGATE_END = 0xdbff
 
-const LOW_SURROGATE_START = 0xDC00
+const LOW_SURROGATE_START = 0xdc00
 
-const REGIONAL_INDICATOR_START = 0x1F1E6
-const REGIONAL_INDICATOR_END = 0x1F1FF
+const REGIONAL_INDICATOR_START = 0x1f1e6
+const REGIONAL_INDICATOR_END = 0x1f1ff
 
 const FITZPATRICK_MODIFIER_START = 0x1f3fb
 const FITZPATRICK_MODIFIER_END = 0x1f3ff
 
-const VARIATION_MODIFIER_START = 0xFE00
-const VARIATION_MODIFIER_END = 0xFE0F
+const VARIATION_MODIFIER_START = 0xfe00
+const VARIATION_MODIFIER_END = 0xfe0f
+
+const ZWJ = 0x200d
 
 function runes (string) {
   if (typeof string !== 'string') {
     throw new Error('string cannot be undefined or null')
   }
   const result = []
   let i = 0
-  let increment
+  let increment = 0
   while (i < string.length) {
-    increment = nextUnits(i, string)
+    increment += nextUnits(i + increment, string)
+    if (isZeroWidthJoiner(string[i + increment])) {
+      increment++
+      continue
+    }
     result.push(string.substring(i, i + increment))
     i += increment
+    increment = 0
   }
   return result
 }
@@ -37,7 +44,6 @@ function runes (string) {
 // Variations: 2 code units
 function nextUnits (i, string) {
   const current = string[i]
-
   // If we have variation selector at next position, we can handle it as pair
   if (isVariationSelector(string[i + 1])) {
     return 2
@@ -90,6 +96,10 @@ function isVariationSelector (string) {
   return typeof string === 'string' && betweenInclusive(string.charCodeAt(0), VARIATION_MODIFIER_START, VARIATION_MODIFIER_END)
 }
 
+function isZeroWidthJoiner (string) {
+  return typeof string === 'string' && string.charCodeAt(0) === ZWJ
+}
+
 function codePointFromSurrogatePair (pair) {
   const highOffset = pair.charCodeAt(0) - HIGH_SURROGATE_START
   const lowOffset = pair.charCodeAt(1) - LOW_SURROGATE_START

package.json

@@ -1,17 +1,18 @@
 {
   "name": "runes",
-  "version": "0.1.0",
+  "version": "0.2.0",
   "description": "Unicode-aware JS string splitting",
   "main": "index.js",
   "scripts": {
-    "test": "eslint . && mocha test"
+    "test": "eslint . && ava test"
   },
   "repository": {
     "type": "git",
     "url": "git+ssh://[email protected]/dotcypress/runes.git"
   },
   "keywords": [
     "unicode",
+    "emoji",
     "runes",
     "split",
     "split string"
@@ -29,12 +30,12 @@
     "index.js"
   ],
   "devDependencies": {
+    "ava": "^0.16.0",
     "eslint": "^3.3.1",
-    "eslint-config-standard": "^5.3.1",
-    "eslint-plugin-mocha": "^4.3.0",
+    "eslint-config-standard": "^6.0.0",
+    "eslint-plugin-ava": "^3.0.0",
     "eslint-plugin-promise": "^2.0.1",
     "eslint-plugin-standard": "^2.0.0",
-    "mocha": "^3.0.2",
     "should": "^11.1.0"
   }
 }

readme.md

@@ -2,7 +2,7 @@
 [![Build Status](https://img.shields.io/travis/dotcypress/runes.svg?branch=master&style=flat-square)](https://travis-ci.org/dotcypress/runes)
 [![NPM Version](https://img.shields.io/npm/v/runes.svg?style=flat-square)](https://www.npmjs.com/package/runes)
 
-Unicode-aware JS string splitting
+Unicode-aware JS string splitting with full Emoji support.
 
 Split a string into its constituent characters, without munging emoji and other non-BMP code points.
 
@@ -19,13 +19,22 @@ $ npm install runes
 ## Example
 
 ```js
-
 const runes = require('runes')
 
-const example = 'Emoji 🤖'
+// Standard String.split
+'♥️'.split('') => ['♥', '️']
+'Emoji 🤖'.split('') => ['E', 'm', 'o', 'j', 'i', ' ', '�', '�']
+'👩‍👩‍👧‍👦'.split('') => ['�', '�', '‍', '�', '�', '‍', '�', '�', '‍', '�', '�']
+
+// ES6 string iterator
+[...'♥️'] => [ '♥', '️' ]
+[...'Emoji 🤖'] => [ 'E', 'm', 'o', 'j', 'i', ' ', '🤖' ]
+[...'👩‍👩‍👧‍👦'] => [ '👩', '', '👩', '', '👧', '', '👦' ]
 
-example.split('') // ["E", "m", "o", "j", "i", " ", "�", "�"]
-runes(example)    // ["E", "m", "o", "j", "i", " ", "🤖"]
+// Runes
+runes('♥️') => ['♥️']
+runes('Emoji 🤖') => ['E', 'm', 'o', 'j', 'i', ' ', '🤖']
+runes('👩‍👩‍👧‍👦') => ['👩‍👩‍👧‍👦']
 
 ```
 

test/index.js

@@ -0,0 +1,105 @@
+'use strict'
+
+require('should')
+const test = require('ava')
+
+const runes = require('../')
+
+test('✂️  Runes should handle emoji in middle', (t) => {
+  runes('abc😤def').should.be.deepEqual(['a', 'b', 'c', '😤', 'd', 'e', 'f'])
+})
+
+test('✂️  Runes should handle leading emoji', (t) => {
+  runes('🍕abd').should.be.deepEqual(['🍕', 'a', 'b', 'd'])
+})
+
+test('✂️  Runes should handle emoji on end', (t) => {
+  runes('123🍥').should.be.deepEqual(['1', '2', '3', '🍥'])
+})
+
+test('✂️  Runes should handle emoji', (t) => {
+  runes('🍕⚽⛵✨⏳☕⏰🇯🇲😍👍💅😋👭👯✊👸🏽❤️').should.be.deepEqual([
+    '🍕', '⚽', '⛵', '✨', '⏳', '☕', '⏰', '🇯🇲',
+    '😍', '👍', '💅', '😋', '👭', '👯', '✊', '👸🏽', '❤️'
+  ])
+})
+
+test('✂️  Runes should handle checkmark', (t) => {
+  runes('123🍕✓').should.be.deepEqual(['1', '2', '3', '🍕', '✓'])
+})
+
+test('✂️  Runes should handle ZERO WIDTH JOINER', (t) => {
+  runes('👨‍👩‍👧').should.be.deepEqual(['👨‍👩‍👧'])
+})
+
+test('✂️  Runes should handle ZERO WIDTH JOINER', (t) => {
+  runes('👨‍👨‍👧‍👧').should.be.deepEqual(['👨‍👨‍👧‍👧'])
+})
+
+test('✂️  Runes should reverse', (t) => {
+  const reversed = runes('123🍕✓').reverse().join('')
+  const contReversed = runes(reversed).reverse().join('')
+  reversed.should.equal('✓🍕321')
+  contReversed.should.equal('123🍕✓')
+})
+
+test('✂️  Runes should handle single char', (t) => {
+  runes('a').should.be.deepEqual(['a'])
+})
+
+test('✂️  Runes should handle regular string', (t) => {
+  runes('Hello').should.be.deepEqual(['H', 'e', 'l', 'l', 'o'])
+})
+
+test('✂️  Runes should handle chinese', (t) => {
+  const string = '𨭎", "𠬠", and "𩷶"'
+  const result = runes(string)
+  result.length.should.equal(16)
+  result[0].should.equal('𨭎')
+  result[1].should.equal('"')
+  result[5].should.equal('𠬠')
+  result[6].should.equal('"')
+  result[14].should.equal('𩷶')
+  result[15].should.equal('"')
+})
+
+test('✂️  Runes should handle math script', (t) => {
+  runes('𝒞𝒯𝒮𝒟').should.be.deepEqual(['𝒞', '𝒯', '𝒮', '𝒟'])
+})
+
+test('✂️  Runes should handle fraktur', (t) => {
+  runes('𝔅𝔎').should.be.deepEqual(['𝔅', '𝔎'])
+})
+
+test('✂️  Runes should handle acrophonic', (t) => {
+  const string = '𐅧, 𐅨, and 𐅩'
+  const result = runes(string)
+  result.length.should.equal(11)
+  result[0].should.equal('𐅧')
+  result[1].should.equal(',')
+  result[3].should.equal('𐅨')
+  result[4].should.equal(',')
+  result[10].should.equal('𐅩')
+})
+
+test('✂️  Runes should handle arabic', (t) => {
+  runes('ځڂڃڄڅچڇڈ').should.be.deepEqual(['ځ', 'ڂ', 'ڃ', 'ڄ', 'څ', 'چ', 'ڇ', 'ڈ'])
+})
+
+test('✂️  Runes should handle skin tone indicators', (t) => {
+  runes('🎅🏻🎅🏼🎅🏽🎅🏾🎅🏿').should.be.deepEqual(['🎅🏻', '🎅🏼', '🎅🏽', '🎅🏾', '🎅🏿'])
+})
+
+test('✂️  Runes should handle country flags/regional indicator characters', (t) => {
+  runes('🇦🇸').should.be.deepEqual(['🇦🇸'])
+})
+
+test('✂️  Runes should handle empty string', (t) => {
+  runes('').should.be.deepEqual([])
+})
+
+test('✂️  Runes should throw for null and undefined', (t) => {
+  (function () {
+    runes()
+  }).should.throw()
+})