Merge pull request #148 from nonara/fix-144

Fix 144
taoqf · Sep 18, 2021 · fd3ec55 · fd3ec55
2 parents df01360 + 5a44c8f
commit fd3ec55
Show file tree

Hide file tree

Showing 2 changed files with 33 additions and 11 deletions.
diff --git a/src/nodes/html.ts b/src/nodes/html.ts
@@ -1025,6 +1025,7 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 	let currentParent = root;
 	const stack = [root];
 	let lastTextPos = -1;
+	let noNestedTagIndex: undefined | number = undefined;
 	let match: RegExpExecArray;
 	// https://github.com/taoqf/node-html-parser/issues/38
 	data = `<${frameflag}>${data}</${frameflag}>`;
@@ -1081,9 +1082,13 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 				}
 			}
 
-			if (currentParent.rawTagName === 'a' && match[2] === 'a') {
-				stack.pop();
-				currentParent = arr_back(stack);
+			// Prevent nested A tags by terminating the last A and starting a new one : see issue #144
+			if (match[2] === 'a' || match[2] === 'A') {
+				if (noNestedTagIndex !== undefined) {
+					stack.splice(noNestedTagIndex);
+					currentParent = arr_back(stack);
+				}
+				noNestedTagIndex = stack.length;
 			}
 
 			const tagEndPos = kMarkupPattern.lastIndex;
@@ -1123,6 +1128,7 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 		// Handle closing tags or self-closed elements (ie </tag> or <br>)
 		if (match[1] || match[4] || kSelfClosingElements[match[2]]) {
 			while (true) {
+				if (match[2] === 'a' || match[2] === 'A') noNestedTagIndex = undefined;
 				if (currentParent.rawTagName === match[2]) {
 					// Update range end for closed tag
 					(<[number, number]>currentParent.range)[1] = createRange(-1, Math.max(lastTextPos, tagEndPos))[1];

diff --git a/test/144.js b/test/144.js
@@ -1,18 +1,34 @@
-const { parse } = require('../dist');
+const { parse, NodeType } = require('../dist');
 
+// Also see comments on https://github.com/taoqf/node-html-parser/pull/148 for additional issues corrected
 describe('issue 144', function () {
 	it('Nested A tags parsed improperly', function () {
-		const html = `<a href="#">link <a href="#">nested link</a> end</a>`;
+		const html = `<A href="#"><b>link <a href="#">nested link</a> end</b></A>`;
+
 		const root = parse(html);
-		root.innerHTML.should.eql(`<a href="#">link </a><a href="#">nested link</a> end`);
+
+		root.innerHTML.should.eql(`<A href="#"><b>link </b></A><a href="#">nested link</a> end`);
 		root.childNodes.length.should.eql(3);
+
 		const a1 = root.childNodes[0];
 		a1.tagName.should.eql('A');
-		a1.nodeType.should.eql(1);
+		a1.nodeType.should.eql(NodeType.ELEMENT_NODE);
+		a1.childNodes.length.should.eql(1);
+
+		const b = a1.childNodes[0];
+		b.tagName.should.eql('B');
+		b.childNodes.length.should.eql(1);
+		b.text.should.eql('link ');
+
 		const a2 = root.childNodes[1];
-		a2.nodeType.should.eql(1);
-		const t1 = root.childNodes[2];
-		t1.nodeType.should.eql(3);
-		t1.textContent.should.eql(' end');
+		a2.tagName.should.eql('A');
+		a2.nodeType.should.eql(NodeType.ELEMENT_NODE);
+		a2.childNodes.length.should.eql(1);
+		a2.childNodes[0].nodeType.should.eql(NodeType.TEXT_NODE);
+		a2.text.should.eql('nested link');
+
+		const endText = root.childNodes[2];
+		endText.nodeType.should.eql(NodeType.TEXT_NODE);
+		endText.textContent.should.eql(' end');
 	});
 });