Upgrade to Unicode 9.0.0 (closes #171)

slevithan · Apr 16, 2017 · 5b4ec78 · 5b4ec78
1 parent 60ae9ec
commit 5b4ec78
Show file tree

Hide file tree

Showing 7 changed files with 275 additions and 135 deletions.
diff --git a/README.md b/README.md
@@ -116,7 +116,7 @@ XRegExp('^\\pS$').test('💩'); // -> true
 
 Opting in to astral mode disables the use of `\p{…}` and `\P{…}` within character classes. In astral mode, use e.g. `(\pL|[0-9_])+` instead of `[\pL0-9_]+`.
 
-XRegExp uses Unicode 8.0.0.
+XRegExp uses Unicode 9.0.0.
 
 ### XRegExp.build
 

diff --git a/src/addons/unicode-blocks.js b/src/addons/unicode-blocks.js
@@ -13,7 +13,7 @@ module.exports = function(XRegExp) {
      * `\p{InBasicLatin}`. Token names are case insensitive, and any spaces, hyphens, and
      * underscores are ignored.
      *
-     * Uses Unicode 8.0.0.
+     * Uses Unicode 9.0.0.
      *
      * @requires XRegExp, Unicode Base
      */
@@ -23,6 +23,10 @@ module.exports = function(XRegExp) {
     }
 
     XRegExp.addUnicodeData([
+        {
+            name: 'InAdlam',
+            astral: '\uD83A[\uDD00-\uDD5F]'
+        },
         {
             name: 'InAegean_Numbers',
             astral: '\uD800[\uDD00-\uDD3F]'
@@ -119,6 +123,10 @@ module.exports = function(XRegExp) {
             name: 'InBengali',
             bmp: '\u0980-\u09FF'
         },
+        {
+            name: 'InBhaiksuki',
+            astral: '\uD807[\uDC00-\uDC6F]'
+        },
         {
             name: 'InBlock_Elements',
             bmp: '\u2580-\u259F'
@@ -197,15 +205,15 @@ module.exports = function(XRegExp) {
         },
         {
             name: 'InCJK_Unified_Ideographs_Extension_C',
-            astral: '\uD86D[\uDC00-\uDF3F]|[\uD86A-\uD86C][\uDC00-\uDFFF]|\uD869[\uDF00-\uDFFF]'
+            astral: '\uD869[\uDF00-\uDFFF]|[\uD86A-\uD86C][\uDC00-\uDFFF]|\uD86D[\uDC00-\uDF3F]'
         },
         {
             name: 'InCJK_Unified_Ideographs_Extension_D',
             astral: '\uD86D[\uDF40-\uDFFF]|\uD86E[\uDC00-\uDC1F]'
         },
         {
             name: 'InCJK_Unified_Ideographs_Extension_E',
-            astral: '[\uD86F-\uD872][\uDC00-\uDFFF]|\uD873[\uDC00-\uDEAF]|\uD86E[\uDC20-\uDFFF]'
+            astral: '\uD86E[\uDC20-\uDFFF]|[\uD86F-\uD872][\uDC00-\uDFFF]|\uD873[\uDC00-\uDEAF]'
         },
         {
             name: 'InCarian',
@@ -299,6 +307,10 @@ module.exports = function(XRegExp) {
             name: 'InCyrillic_Extended_B',
             bmp: '\uA640-\uA69F'
         },
+        {
+            name: 'InCyrillic_Extended_C',
+            bmp: '\u1C80-\u1C8F'
+        },
         {
             name: 'InCyrillic_Supplement',
             bmp: '\u0500-\u052F'
@@ -399,6 +411,10 @@ module.exports = function(XRegExp) {
             name: 'InGlagolitic',
             bmp: '\u2C00-\u2C5F'
         },
+        {
+            name: 'InGlagolitic_Supplement',
+            astral: '\uD838[\uDC00-\uDC2F]'
+        },
         {
             name: 'InGothic',
             astral: '\uD800[\uDF30-\uDF4F]'
@@ -479,6 +495,10 @@ module.exports = function(XRegExp) {
             name: 'InIdeographic_Description_Characters',
             bmp: '\u2FF0-\u2FFF'
         },
+        {
+            name: 'InIdeographic_Symbols_and_Punctuation',
+            astral: '\uD81B[\uDFE0-\uDFFF]'
+        },
         {
             name: 'InImperial_Aramaic',
             astral: '\uD802[\uDC40-\uDC5F]'
@@ -639,6 +659,10 @@ module.exports = function(XRegExp) {
             name: 'InManichaean',
             astral: '\uD802[\uDEC0-\uDEFF]'
         },
+        {
+            name: 'InMarchen',
+            astral: '\uD807[\uDC70-\uDCBF]'
+        },
         {
             name: 'InMathematical_Alphanumeric_Symbols',
             astral: '\uD835[\uDC00-\uDFFF]'
@@ -689,7 +713,7 @@ module.exports = function(XRegExp) {
         },
         {
             name: 'InMiscellaneous_Symbols_and_Pictographs',
-            astral: '\uD83D[\uDC00-\uDDFF]|\uD83C[\uDF00-\uDFFF]'
+            astral: '\uD83C[\uDF00-\uDFFF]|\uD83D[\uDC00-\uDDFF]'
         },
         {
             name: 'InMiscellaneous_Technical',
@@ -707,6 +731,10 @@ module.exports = function(XRegExp) {
             name: 'InMongolian',
             bmp: '\u1800-\u18AF'
         },
+        {
+            name: 'InMongolian_Supplement',
+            astral: '\uD805[\uDE60-\uDE7F]'
+        },
         {
             name: 'InMro',
             astral: '\uD81A[\uDE40-\uDE6F]'
@@ -743,6 +771,10 @@ module.exports = function(XRegExp) {
             name: 'InNew_Tai_Lue',
             bmp: '\u1980-\u19DF'
         },
+        {
+            name: 'InNewa',
+            astral: '\uD805[\uDC00-\uDC7F]'
+        },
         {
             name: 'InNumber_Forms',
             bmp: '\u2150-\u218F'
@@ -795,6 +827,10 @@ module.exports = function(XRegExp) {
             name: 'InOrnamental_Dingbats',
             astral: '\uD83D[\uDE50-\uDE7F]'
         },
+        {
+            name: 'InOsage',
+            astral: '\uD801[\uDCB0-\uDCFF]'
+        },
         {
             name: 'InOsmanya',
             astral: '\uD801[\uDC80-\uDCAF]'
@@ -995,6 +1031,14 @@ module.exports = function(XRegExp) {
             name: 'InTamil',
             bmp: '\u0B80-\u0BFF'
         },
+        {
+            name: 'InTangut',
+            astral: '[\uD81C-\uD821][\uDC00-\uDFFF]'
+        },
+        {
+            name: 'InTangut_Components',
+            astral: '\uD822[\uDC00-\uDEFF]'
+        },
         {
             name: 'InTelugu',
             bmp: '\u0C00-\u0C7F'