Merge pull request #1304 from capricorn86/1039-can-not-parse-correctl…

…y-html-with-nested-ul-and-li-tags 1039 can not parse correctly html with nested ul and li tags
capricorn86 · Mar 13, 2024 · 713aa3c · 713aa3c
2 parents 4970c69 + 23e9616
commit 713aa3c
Show file tree

Hide file tree

Showing 11 changed files with 815 additions and 200 deletions.
diff --git a/packages/happy-dom/src/config/HTMLElementConfig.ts b/packages/happy-dom/src/config/HTMLElementConfig.ts
diff --git a/packages/happy-dom/src/config/HTMLElementConfigContentModelEnum.ts b/packages/happy-dom/src/config/HTMLElementConfigContentModelEnum.ts
@@ -0,0 +1,9 @@
+enum HTMLElementConfigContentModelEnum {
+	rawText = 'rawText',
+	noSelfDescendants = 'noSelfDescendants',
+	noFirsLevelSelfDescendants = 'noFirsLevelSelfDescendants',
+	noDescendants = 'noDescendants',
+	anyDescendants = 'anyDescendants'
+}
+
+export default HTMLElementConfigContentModelEnum;
diff --git a/packages/happy-dom/src/config/HTMLElementLocalNameToClass.ts b/packages/happy-dom/src/config/HTMLElementLocalNameToClass.ts
diff --git a/packages/happy-dom/src/config/HTMLElementPlainText.ts b/packages/happy-dom/src/config/HTMLElementPlainText.ts
diff --git a/packages/happy-dom/src/config/HTMLElementUnnestable.ts b/packages/happy-dom/src/config/HTMLElementUnnestable.ts
diff --git a/packages/happy-dom/src/config/HTMLElementVoid.ts b/packages/happy-dom/src/config/HTMLElementVoid.ts
diff --git a/packages/happy-dom/src/config/IHTMLElementConfigEntity.ts b/packages/happy-dom/src/config/IHTMLElementConfigEntity.ts
@@ -0,0 +1,8 @@
+import HTMLElementConfigContentModelEnum from './HTMLElementConfigContentModelEnum.js';
+
+export default interface IHTMLElementConfigEntity {
+	className: string;
+	localName: string;
+	tagName: string;
+	contentModel: HTMLElementConfigContentModelEnum;
+}
diff --git a/packages/happy-dom/src/nodes/document/Document.ts b/packages/happy-dom/src/nodes/document/Document.ts
@@ -8,7 +8,6 @@ import DocumentFragment from '../document-fragment/DocumentFragment.js';
 import XMLParser from '../../xml-parser/XMLParser.js';
 import Event from '../../event/Event.js';
 import DOMImplementation from '../../dom-implementation/DOMImplementation.js';
-import HTMLElementLocalNameToClass from '../../config/HTMLElementLocalNameToClass.js';
 import INodeFilter from '../../tree-walker/INodeFilter.js';
 import NamespaceURI from '../../config/NamespaceURI.js';
 import DocumentType from '../document-type/DocumentType.js';
@@ -51,6 +50,7 @@ import ISVGElementTagNameMap from '../../config/ISVGElementTagNameMap.js';
 import ISVGElement from '../svg-element/ISVGElement.js';
 import IHTMLFormElement from '../html-form-element/IHTMLFormElement.js';
 import IHTMLAnchorElement from '../html-anchor-element/IHTMLAnchorElement.js';
+import HTMLElementConfig from '../../config/HTMLElementConfig.js';
 
 const PROCESSING_INSTRUCTION_TARGET_REGEXP = /^[a-z][a-z0-9-]+$/;
 
@@ -1131,7 +1131,9 @@ export default class Document extends Node implements IDocument {
 		}
 
 		const localName = qualifiedName.toLowerCase();
-		const elementClass = this[PropertySymbol.ownerWindow][HTMLElementLocalNameToClass[localName]];
+		const elementClass = HTMLElementConfig[localName]
+			? this[PropertySymbol.ownerWindow][HTMLElementConfig[localName].className]
+			: null;
 
 		// Known HTML element
 		if (elementClass) {

diff --git a/packages/happy-dom/src/xml-parser/XMLParser.ts b/packages/happy-dom/src/xml-parser/XMLParser.ts
@@ -1,16 +1,15 @@
 import IDocument from '../nodes/document/IDocument.js';
 import * as PropertySymbol from '../PropertySymbol.js';
-import HTMLElementVoid from '../config/HTMLElementVoid.js';
-import HTMLElementUnnestable from '../config/HTMLElementUnnestable.js';
 import NamespaceURI from '../config/NamespaceURI.js';
 import HTMLScriptElement from '../nodes/html-script-element/HTMLScriptElement.js';
 import IElement from '../nodes/element/IElement.js';
 import HTMLLinkElement from '../nodes/html-link-element/HTMLLinkElement.js';
-import HTMLElementPlainText from '../config/HTMLElementPlainText.js';
 import IDocumentType from '../nodes/document-type/IDocumentType.js';
 import INode from '../nodes/node/INode.js';
 import IDocumentFragment from '../nodes/document-fragment/IDocumentFragment.js';
+import HTMLElementConfig from '../config/HTMLElementConfig.js';
 import * as Entities from 'entities';
+import HTMLElementConfigContentModelEnum from '../config/HTMLElementConfigContentModelEnum.js';
 
 /**
  * Markup RegExp.
@@ -58,6 +57,8 @@ const DOCUMENT_TYPE_ATTRIBUTE_REGEXP = /"([^"]+)"/gm;
 
 /**
  * XML parser.
+ *
+ * @see https://html.spec.whatwg.org/multipage/indices.html
  */
 export default class XMLParser {
 	/**
@@ -77,12 +78,11 @@ export default class XMLParser {
 	): IElement | IDocumentFragment | IDocument {
 		const root = options && options.rootNode ? options.rootNode : document.createDocumentFragment();
 		const stack: INode[] = [root];
+		const stackTagNames: string[] = [];
 		const markupRegexp = new RegExp(MARKUP_REGEXP, 'gm');
 		const { evaluateScripts = false } = options || {};
-		const unnestableTagNames: string[] = [];
 		let currentNode: INode | null = root;
 		let match: RegExpExecArray;
-		let plainTextTagName: string | null = null;
 		let readState: MarkupReadStateEnum = MarkupReadStateEnum.startOrEndTag;
 		let startTagIndex = 0;
 		let lastIndex = 0;
@@ -108,19 +108,31 @@ export default class XMLParser {
 							// Start tag.
 							const tagName = match[1].toUpperCase();
 							const localName = tagName === 'SVG' ? 'svg' : match[1];
+							const config = HTMLElementConfig[localName];
 
 							// Some elements are not allowed to be nested (e.g. "<a><a></a></a>" is not allowed.).
 							// Therefore we need to auto-close the tag, so that it become valid (e.g. "<a></a><a></a>").
-							const unnestableTagNameIndex = unnestableTagNames.indexOf(tagName);
-							if (unnestableTagNameIndex !== -1) {
-								unnestableTagNames.splice(unnestableTagNameIndex, 1);
+							if (
+								config?.contentModel ===
+									HTMLElementConfigContentModelEnum.noFirsLevelSelfDescendants &&
+								stackTagNames[stackTagNames.length - 1] === tagName
+							) {
+								stack.pop();
+								stackTagNames.pop();
+								currentNode = stack[stack.length - 1] || root;
+							} else if (
+								config?.contentModel === HTMLElementConfigContentModelEnum.noSelfDescendants &&
+								stackTagNames.includes(tagName)
+							) {
 								while (currentNode !== root) {
 									if ((<IElement>currentNode)[PropertySymbol.tagName].toUpperCase() === tagName) {
 										stack.pop();
+										stackTagNames.pop();
 										currentNode = stack[stack.length - 1] || root;
 										break;
 									}
 									stack.pop();
+									stackTagNames.pop();
 									currentNode = stack[stack.length - 1] || root;
 								}
 							}
@@ -136,25 +148,18 @@ export default class XMLParser {
 							currentNode.appendChild(newElement);
 							currentNode = newElement;
 							stack.push(currentNode);
+							stackTagNames.push(tagName);
 							readState = MarkupReadStateEnum.insideStartTag;
 							startTagIndex = markupRegexp.lastIndex;
 						} else if (match[2]) {
 							// End tag.
 
 							if (
 								match[2].toUpperCase() ===
-								(<IElement>currentNode)[PropertySymbol.tagName].toUpperCase()
+								(<IElement>currentNode)[PropertySymbol.tagName]?.toUpperCase()
 							) {
-								// Some elements are not allowed to be nested (e.g. "<a><a></a></a>" is not allowed.).
-								// Therefore we need to auto-close the tag, so that it become valid (e.g. "<a></a><a></a>").
-								const unnestableTagNameIndex = unnestableTagNames.indexOf(
-									(<IElement>currentNode)[PropertySymbol.tagName].toUpperCase()
-								);
-								if (unnestableTagNameIndex !== -1) {
-									unnestableTagNames.splice(unnestableTagNameIndex, 1);
-								}
-
 								stack.pop();
+								stackTagNames.pop();
 								currentNode = stack[stack.length - 1] || root;
 							}
 						} else if (
@@ -201,8 +206,6 @@ export default class XMLParser {
 					case MarkupReadStateEnum.insideStartTag:
 						// End of start tag
 						if (match[7] || match[8]) {
-							// End of start tag.
-
 							// Attribute name and value.
 
 							const attributeString = xml.substring(startTagIndex, match.index);
@@ -257,33 +260,27 @@ export default class XMLParser {
 							// We need to check if the attribute string is read completely.
 							// The attribute string can potentially contain "/>" or ">".
 							if (hasAttributeStringEnded) {
+								const config = HTMLElementConfig[(<IElement>currentNode)[PropertySymbol.localName]];
+
 								// Checks if the tag is a self closing tag (ends with "/>") or void element.
 								// When it is a self closing tag or void element it should be closed immediately.
 								// Self closing tags are not allowed in the HTML namespace, but the parser should still allow it for void elements.
 								// Self closing tags is supported in the SVG namespace.
 								if (
-									HTMLElementVoid[(<IElement>currentNode)[PropertySymbol.tagName]] ||
+									config?.contentModel === HTMLElementConfigContentModelEnum.noDescendants ||
+									// SVG tag is self closing (<svg/>).
 									(match[7] &&
 										(<IElement>currentNode)[PropertySymbol.namespaceURI] === NamespaceURI.svg)
 								) {
 									stack.pop();
+									stackTagNames.pop();
 									currentNode = stack[stack.length - 1] || root;
 									readState = MarkupReadStateEnum.startOrEndTag;
 								} else {
-									// Plain text elements such as <script> and <style> should only contain text.
-									plainTextTagName = HTMLElementPlainText[
-										(<IElement>currentNode)[PropertySymbol.tagName]
-									]
-										? (<IElement>currentNode)[PropertySymbol.tagName]
-										: null;
-
-									readState = !!plainTextTagName
-										? MarkupReadStateEnum.plainTextContent
-										: MarkupReadStateEnum.startOrEndTag;
-
-									if (HTMLElementUnnestable[(<IElement>currentNode)[PropertySymbol.tagName]]) {
-										unnestableTagNames.push((<IElement>currentNode)[PropertySymbol.tagName]);
-									}
+									readState =
+										config?.contentModel === HTMLElementConfigContentModelEnum.rawText
+											? MarkupReadStateEnum.plainTextContent
+											: MarkupReadStateEnum.startOrEndTag;
 								}
 
 								startTagIndex = markupRegexp.lastIndex;
@@ -292,15 +289,17 @@ export default class XMLParser {
 
 						break;
 					case MarkupReadStateEnum.plainTextContent:
-						if (match[2] && match[2].toUpperCase() === plainTextTagName) {
+						const tagName = currentNode[PropertySymbol.tagName];
+
+						if (tagName && match[2] && match[2].toUpperCase() === tagName) {
 							// End of plain text tag.
 
 							// Scripts are not allowed to be executed when they are parsed using innerHTML, outerHTML, replaceWith() etc.
 							// However, they are allowed to be executed when document.write() is used.
 							// See: https://developer.mozilla.org/en-US/docs/Web/API/HTMLScriptElement
-							if (plainTextTagName === 'SCRIPT') {
+							if (tagName === 'SCRIPT') {
 								(<HTMLScriptElement>currentNode)[PropertySymbol.evaluateScript] = evaluateScripts;
-							} else if (plainTextTagName === 'LINK') {
+							} else if (tagName === 'LINK') {
 								// An assumption that the same rule should be applied for the HTMLLinkElement is made here.
 								(<HTMLLinkElement>currentNode)[PropertySymbol.evaluateCSS] = evaluateScripts;
 							}
@@ -313,8 +312,8 @@ export default class XMLParser {
 							);
 
 							stack.pop();
+							stackTagNames.pop();
 							currentNode = stack[stack.length - 1] || root;
-							plainTextTagName = null;
 							readState = MarkupReadStateEnum.startOrEndTag;
 						}