JS: Added support for new regex syntax (#3399)

components/prism-javascript.js

@@ -53,8 +53,24 @@ Prism.languages.javascript['class-name'][0].pattern = /(\b(?:class|extends|imple
 
 Prism.languages.insertBefore('javascript', 'keyword', {
 	'regex': {
-		// eslint-disable-next-line regexp/no-dupe-characters-character-class
-		pattern: /((?:^|[^$\w\xA0-\uFFFF."'\])\s]|\b(?:return|yield))\s*)\/(?:\[(?:[^\]\\\r\n]|\\.)*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}(?=(?:\s|\/\*(?:[^*]|\*(?!\/))*\*\/)*(?:$|[\r\n,.;:})\]]|\/\/))/,
+		pattern: RegExp(
+			// lookbehind
+			// eslint-disable-next-line regexp/no-dupe-characters-character-class
+			/((?:^|[^$\w\xA0-\uFFFF."'\])\s]|\b(?:return|yield))\s*)/.source +
+			// Regex pattern:
+			// There are 2 regex patterns here. The RegExp set notation proposal added support for nested character
+			// classes if the `v` flag is present. Unfortunately, nested CCs are both context-free and incompatible
+			// with the only syntax, so we have to define 2 different regex patterns.
+			/\//.source +
+			'(?:' +
+			/(?:\[(?:[^\]\\\r\n]|\\.)*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}/.source +
+			'|' +
+			// `v` flag syntax. This supports 3 levels of nested character classes.
+			/(?:\[(?:[^[\]\\\r\n]|\\.|\[(?:[^[\]\\\r\n]|\\.|\[(?:[^[\]\\\r\n]|\\.)*\])*\])*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}v[dgimyus]{0,7}/.source +
+			')' +
+			// lookahead
+			/(?=(?:\s|\/\*(?:[^*]|\*(?!\/))*\*\/)*(?:$|[\r\n,.;:})\]]|\/\/))/.source
+		),
 		lookbehind: true,
 		greedy: true,
 		inside: {

prism.js

@@ -1623,8 +1623,24 @@ Prism.languages.javascript['class-name'][0].pattern = /(\b(?:class|extends|imple
 
 Prism.languages.insertBefore('javascript', 'keyword', {
 	'regex': {
-		// eslint-disable-next-line regexp/no-dupe-characters-character-class
-		pattern: /((?:^|[^$\w\xA0-\uFFFF."'\])\s]|\b(?:return|yield))\s*)\/(?:\[(?:[^\]\\\r\n]|\\.)*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}(?=(?:\s|\/\*(?:[^*]|\*(?!\/))*\*\/)*(?:$|[\r\n,.;:})\]]|\/\/))/,
+		pattern: RegExp(
+			// lookbehind
+			// eslint-disable-next-line regexp/no-dupe-characters-character-class
+			/((?:^|[^$\w\xA0-\uFFFF."'\])\s]|\b(?:return|yield))\s*)/.source +
+			// Regex pattern:
+			// There are 2 regex patterns here. The RegExp set notation proposal added support for nested character
+			// classes if the `v` flag is present. Unfortunately, nested CCs are both context-free and incompatible
+			// with the only syntax, so we have to define 2 different regex patterns.
+			/\//.source +
+			'(?:' +
+			/(?:\[(?:[^\]\\\r\n]|\\.)*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}/.source +
+			'|' +
+			// `v` flag syntax. This supports 3 levels of nested character classes.
+			/(?:\[(?:[^[\]\\\r\n]|\\.|\[(?:[^[\]\\\r\n]|\\.|\[(?:[^[\]\\\r\n]|\\.)*\])*\])*\]|\\.|[^/\\\[\r\n])+\/[dgimyus]{0,7}v[dgimyus]{0,7}/.source +
+			')' +
+			// lookahead
+			/(?=(?:\s|\/\*(?:[^*]|\*(?!\/))*\*\/)*(?:$|[\r\n,.;:})\]]|\/\/))/.source
+		),
 		lookbehind: true,
 		greedy: true,
 		inside: {

tests/languages/javascript/regex_feature.test

@@ -9,6 +9,12 @@
 / '1' '2' '3' '4' '5' /
 [/foo/]
 
+// RegExp set notation
+;/[A--B]/v;
+/[A--[0-9]]/v;
+/[\p{Decimal_Number}--[0-9]]/v;
+/[\p{Script=Khmer}&&[\p{Letter}\p{Mark}\p{Number}]]/v;
+
 let a = /regex/m // comment
 let b = condition ? /regex/ : /another one/
 return /regex/;
@@ -88,6 +94,41 @@ yield /regex/;
 	]],
 	["punctuation", "]"],
 
+	["comment", "// RegExp set notation"],
+
+	["punctuation", ";"],
+	["regex", [
+		["regex-delimiter", "/"],
+		["regex-source", "[A--B]"],
+		["regex-delimiter", "/"],
+		["regex-flags", "v"]
+	]],
+	["punctuation", ";"],
+
+	["regex", [
+		["regex-delimiter", "/"],
+		["regex-source", "[A--[0-9]]"],
+		["regex-delimiter", "/"],
+		["regex-flags", "v"]
+	]],
+	["punctuation", ";"],
+
+	["regex", [
+		["regex-delimiter", "/"],
+		["regex-source", "[\\p{Decimal_Number}--[0-9]]"],
+		["regex-delimiter", "/"],
+		["regex-flags", "v"]
+	]],
+	["punctuation", ";"],
+
+	["regex", [
+		["regex-delimiter", "/"],
+		["regex-source", "[\\p{Script=Khmer}&&[\\p{Letter}\\p{Mark}\\p{Number}]]"],
+		["regex-delimiter", "/"],
+		["regex-flags", "v"]
+	]],
+	["punctuation", ";"],
+
 	["keyword", "let"],
 	" a ",
 	["operator", "="],