fix: Preserve invalid nested A tags in AST (see #215 for detail)

Author: nonara

.gitattributes

@@ -0,0 +1,39 @@
+# These settings are for any web project
+
+# Handle line endings automatically for files detected as text
+# and leave all files detected as binary untouched.
+* text=auto
+
+# Force the following filetypes to have unix eols, so Windows does not break them
+*.* text eol=lf
+
+# Windows forced line-endings
+/.idea/* text eol=crlf
+
+#
+## These files are binary and should be left untouched
+#
+
+# (binary is a macro for -text -diff)
+*.png binary
+*.jpg binary
+*.jpeg binary
+*.gif binary
+*.ico binary
+*.mov binary
+*.mp4 binary
+*.mp3 binary
+*.flv binary
+*.fla binary
+*.swf binary
+*.gz binary
+*.zip binary
+*.7z binary
+*.ttf binary
+*.eot binary
+*.woff binary
+*.pyc binary
+*.pdf binary
+*.ez binary
+*.bz2 binary
+*.swp binary

src/nodes/html.ts

@@ -982,6 +982,10 @@ const kElementsClosedByClosing = {
 export interface Options {
 	lowerCaseTagName: boolean;
 	comment: boolean;
+	/**
+	 * @see PR #215 for explanation
+	 */
+	fixNestedATags?: boolean;
 	parseNoneClosedTags?: boolean;
 	blockTextElements: {
 		[tag: string]: boolean;
@@ -1036,7 +1040,7 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 	let match: RegExpExecArray;
 	// https://github.com/taoqf/node-html-parser/issues/38
 	data = `<${frameflag}>${data}</${frameflag}>`;
-	const { lowerCaseTagName } = options;
+	const { lowerCaseTagName, fixNestedATags } = options;
 
 	const dataEndPos = data.length - (frameflag.length + 2);
 	const frameFlagOffset = frameflag.length + 2;
@@ -1097,7 +1101,7 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 			}
 
 			// Prevent nested A tags by terminating the last A and starting a new one : see issue #144
-			if (tagName === 'a' || tagName === 'A') {
+			if (fixNestedATags && (tagName === 'a' || tagName === 'A')) {
 				if (noNestedTagIndex !== undefined) {
 					stack.splice(noNestedTagIndex);
 					currentParent = arr_back(stack);
@@ -1142,7 +1146,7 @@ export function base_parse(data: string, options = { lowerCaseTagName: false, co
 		// Handle closing tags or self-closed elements (ie </tag> or <br>)
 		if (leadingSlash || closingSlash || kSelfClosingElements[tagName]) {
 			while (true) {
-				if (tagName === 'a' || tagName === 'A') noNestedTagIndex = undefined;
+				if (noNestedTagIndex != null && (tagName === 'a' || tagName === 'A')) noNestedTagIndex = undefined;
 				if (currentParent.rawTagName === tagName) {
 					// Update range end for closed tag
 					(<[number, number]>currentParent.range)[1] = createRange(-1, Math.max(lastTextPos, tagEndPos))[1];

test/tests/issues/144.js

@@ -0,0 +1,63 @@
+const { parse, NodeType } = require('@test/test-target');
+
+describe('Nested A Tags', function () {
+	it('Tags preserved by default', function () {
+		const html = `<A href="#"><b>link <a href="#">nested link</a> end</b></A>`;
+
+		const root = parse(html);
+
+		root.innerHTML.should.eql(`<A href="#"><b>link <a href="#">nested link</a> end</b></A>`);
+		root.childNodes.length.should.eql(1);
+
+		const a1 = root.childNodes[0];
+		a1.tagName.should.eql('A');
+		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(3);
+		b.text.should.eql('link nested link end');
+
+		const a2 = b.childNodes[1];
+		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 = b.childNodes[2];
+		endText.nodeType.should.eql(NodeType.TEXT_NODE);
+		endText.textContent.should.eql(' end');
+	});
+
+	it('Tags fixed with fixNestedATags option', function () {
+		const html = `<A href="#"><b>link <a href="#">nested link</a> end</b></A>`;
+
+		const root = parse(html, { fixNestedATags: true });
+
+		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(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.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');
+	});
+});