Haxe: Improved tokenization (#3211)

components.json

@@ -535,6 +535,7 @@
 		"haxe": {
 			"title": "Haxe",
 			"require": "clike",
+			"optional": "regex",
 			"owner": "Golmote"
 		},
 		"hcl": {

components/prism-haxe.js

@@ -1,45 +1,78 @@
 Prism.languages.haxe = Prism.languages.extend('clike', {
-	// Strings can be multi-line
 	'string': {
-		pattern: /(["'])(?:(?!\1)[^\\]|\\[\s\S])*\1/,
+		// Strings can be multi-line
+		pattern: /"(?:[^"\\]|\\[\s\S])*"/,
+		greedy: true
+	},
+	'class-name': [
+		{
+			pattern: /(\b(?:abstract|class|enum|extends|implements|interface|new|typedef)\s+)[A-Z_]\w*/,
+			lookbehind: true,
+		},
+		// based on naming convention
+		/\b[A-Z]\w*/
+	],
+	// The final look-ahead prevents highlighting of keywords if expressions such as "haxe.macro.Expr"
+	'keyword': /\bthis\b|\b(?:abstract|as|break|case|cast|catch|class|continue|default|do|dynamic|else|enum|extends|extern|final|for|from|function|if|implements|import|in|inline|interface|macro|new|null|operator|overload|override|package|private|public|return|static|super|switch|throw|to|try|typedef|untyped|using|var|while)(?!\.)\b/,
+	'function': {
+		pattern: /\b[a-z_]\w*(?=\s*(?:<[^<>]*>\s*)?\()/i,
+		greedy: true
+	},
+	'operator': /\.{3}|\+\+|--|&&|\|\||->|=>|(?:<<?|>{1,3}|[-+*/%!=&|^])=?|[?:~]/
+});
+
+Prism.languages.insertBefore('haxe', 'string', {
+	'string-interpolation': {
+		pattern: /'(?:[^'\\]|\\[\s\S])*'/,
 		greedy: true,
 		inside: {
 			'interpolation': {
-				pattern: /(^|[^\\])\$(?:\w+|\{[^}]+\})/,
+				pattern: /(^|[^\\])\$(?:\w+|\{[^{}]+\})/,
 				lookbehind: true,
 				inside: {
-					'interpolation': {
-						pattern: /^\$\w*/,
-						alias: 'variable'
-					}
-					// See rest below
+					'interpolation-punctuation': {
+						pattern: /^\$\{?|\}$/,
+						alias: 'punctuation'
+					},
+					'expression': {
+						pattern: /[\s\S]+/,
+						inside: Prism.languages.haxe
+					},
 				}
-			}
+			},
+			'string': /[\s\S]+/
 		}
-	},
-	// The final look-ahead prevents highlighting of keywords if expressions such as "haxe.macro.Expr"
-	'keyword': /\bthis\b|\b(?:abstract|as|break|case|cast|catch|class|continue|default|do|dynamic|else|enum|extends|extern|for|from|function|if|implements|import|in|inline|interface|macro|new|null|override|private|public|return|static|super|switch|throw|to|try|typedef|using|var|while)(?!\.)\b/,
-	'operator': /\.{3}|\+\+?|-[->]?|[=!]=?|&&?|\|\|?|<[<=]?|>[>=]?|[*\/%~^]/
+	}
 });
+
 Prism.languages.insertBefore('haxe', 'class-name', {
 	'regex': {
-		pattern: /~\/(?:[^\/\\\r\n]|\\.)+\/[igmsu]*/,
-		greedy: true
+		pattern: /~\/(?:[^\/\\\r\n]|\\.)+\/[a-z]*/,
+		greedy: true,
+		inside: {
+			'regex-flags': /\b[a-z]+$/,
+			'regex-source': {
+				pattern: /^(~\/)[\s\S]+(?=\/$)/,
+				lookbehind: true,
+				alias: 'language-regex',
+				inside: Prism.languages.regex
+			},
+			'regex-delimiter': /^~\/|\/$/,
+		}
 	}
 });
+
 Prism.languages.insertBefore('haxe', 'keyword', {
 	'preprocessor': {
-		pattern: /#\w+/,
-		alias: 'builtin'
+		pattern: /#(?:else|elseif|end|if)\b.*/,
+		alias: 'property'
 	},
 	'metadata': {
-		pattern: /@:?\w+/,
+		pattern: /@:?[\w.]+/,
 		alias: 'symbol'
 	},
 	'reification': {
 		pattern: /\$(?:\w+|(?=\{))/,
-		alias: 'variable'
+		alias: 'important'
 	}
 });
-Prism.languages.haxe['string'].inside['interpolation'].inside.rest = Prism.languages.haxe;
-delete Prism.languages.haxe['class-name'];

tests/languages/haxe!+regex/regex_inclusion.test

@@ -0,0 +1,99 @@
+~/ha\/xe/i
+~/[A-Z0-9._%-]+@[A-Z0-9.-]+.[A-Z][A-Z][A-Z]?/i
+~/(dog|fox)/igmsu
+
+----------------------------------------------------
+
+[
+	["regex", [
+		["regex-delimiter", "~/"],
+		["regex-source", [
+			"ha",
+			["escape", "\\/"],
+			"xe"
+		]],
+		["regex-delimiter", "/"],
+		["regex-flags", "i"]
+	]],
+	["regex", [
+		["regex-delimiter", "~/"],
+		["regex-source", [
+			["char-class", [
+				["char-class-punctuation", "["],
+				["range", [
+					"A",
+					["range-punctuation", "-"],
+					"Z"
+				]],
+				["range", [
+					"0",
+					["range-punctuation", "-"],
+					"9"
+				]],
+				"._%-",
+				["char-class-punctuation", "]"]
+			]],
+			["quantifier", "+"],
+			"@",
+			["char-class", [
+				["char-class-punctuation", "["],
+				["range", [
+					"A",
+					["range-punctuation", "-"],
+					"Z"
+				]],
+				["range", [
+					"0",
+					["range-punctuation", "-"],
+					"9"
+				]],
+				".-",
+				["char-class-punctuation", "]"]
+			]],
+			["quantifier", "+"],
+			["char-set", "."],
+			["char-class", [
+				["char-class-punctuation", "["],
+				["range", [
+					"A",
+					["range-punctuation", "-"],
+					"Z"
+				]],
+				["char-class-punctuation", "]"]
+			]],
+			["char-class", [
+				["char-class-punctuation", "["],
+				["range", [
+					"A",
+					["range-punctuation", "-"],
+					"Z"
+				]],
+				["char-class-punctuation", "]"]
+			]],
+			["char-class", [
+				["char-class-punctuation", "["],
+				["range", [
+					"A",
+					["range-punctuation", "-"],
+					"Z"
+				]],
+				["char-class-punctuation", "]"]
+			]],
+			["quantifier", "?"]
+		]],
+		["regex-delimiter", "/"],
+		["regex-flags", "i"]
+	]],
+	["regex", [
+		["regex-delimiter", "~/"],
+		["regex-source", [
+			["group", ["("]],
+			"dog",
+			["alternation", "|"],
+			"fox",
+			["group", ")"]
+		]],
+		["regex-delimiter", "/"],
+		["regex-flags", "igmsu"]
+	]]
+]