Merge branch 'master' into fix-unclosed

package.json

@@ -1,6 +1,6 @@
 {
 	"name": "node-html-parser",
-	"version": "1.2.14",
+	"version": "1.2.15",
 	"description": "A very fast HTML parser, generating a simplified DOM, with basic element query support.",
 	"main": "dist/index.js",
 	"types": "dist/index.d.ts",

src/nodes/html.ts

@@ -595,6 +595,8 @@ export interface Options {
 	comment?: boolean;
 }
 
+const frameflag = 'documentfragmentcontainer';
+
 /**
  * Parses HTML and returns a root element
  * Parse a chuck of HTML source.
@@ -607,6 +609,8 @@ export function parse(data: string, options = {} as Options) {
 	const stack = [root];
 	let lastTextPos = -1;
 	let match: RegExpExecArray;
+	// https://github.com/taoqf/node-html-parser/issues/38
+	data = `<${frameflag}>${data}</${frameflag}>`;
 	while (match = kMarkupPattern.exec(data)) {
 		if (lastTextPos > -1) {
 			if (lastTextPos + match[0].length < kMarkupPattern.lastIndex) {
@@ -616,6 +620,9 @@ export function parse(data: string, options = {} as Options) {
 			}
 		}
 		lastTextPos = kMarkupPattern.lastIndex;
+		if (match[2] === frameflag) {
+			continue;
+		}
 		if (match[0][1] === '!') {
 			// this is a comment
 			if (options.comment) {
@@ -642,8 +649,9 @@ export function parse(data: string, options = {} as Options) {
 					currentParent = arr_back(stack);
 				}
 			}
-			currentParent = currentParent.appendChild(
-				new HTMLElement(match[2], attrs, match[3]));
+			// ignore container tag we add above
+			// https://github.com/taoqf/node-html-parser/issues/38
+			currentParent = currentParent.appendChild(new HTMLElement(match[2], attrs, match[3]));
 			stack.push(currentParent);
 			if (kBlockTextElements[match[2]]) {
 				// a little test to find next </script> or </style> ...

test/html.js

@@ -429,10 +429,12 @@ describe('HTML Parser', function () {
 		});
 
 		describe('#removeAttribute', function () {
-			const root = parseHTML('<input required>');
-			const input = root.firstChild;
-			input.removeAttribute('required');
-			input.toString().should.eql('<input>');
+			it('should remove attribute required', function () {
+				const root = parseHTML('<input required>');
+				const input = root.firstChild;
+				input.removeAttribute('required');
+				input.toString().should.eql('<input>');
+			});
 		});
 
 		describe('#hasAttribute', function () {

test/parse.js

@@ -0,0 +1,13 @@
+const { parse } = require('../dist');
+
+// https://github.com/taoqf/node-html-parser/issues/38
+describe('HTML Parser', function () {
+	it('should parse text element', function () {
+		const root = parse('foo bar<div>aaa</div>');
+		root.toString().should.eql('foo bar<div>aaa</div>');
+	});
+	it('should parse pure text element', function () {
+		const root = parse('foo bar');
+		root.toString().should.eql('foo bar');
+	});
+})