Throw errors from HTML parsing

Replace some TODOs with actual exception throwing. The exceptions aren’t
great, but at least they’re better than just ignoring the errors.

The first case doesn’t have a test because, as far as I can tell, it’s
basically impossible to trigger. There are apparently very few
situations where loadHTML() will return false [1] rather than trying to
fix up the markup somehow, and we’re already guaranteeing that the HTML
isn’t empty and that the options are valid.

The second case can raise a variety of errors; just test one of them.
Turning the array of LibXMLError instances (which aren’t Throwable) into
a “tree” / “stack” of Exception instances isn’t super nice, but better
than doing nothing or only throwing an Exception for the first error.

Note that we have to ignore / skip one error: libxml may complain about
“invalid” <template> tags, but we need those for SFC handling.

[1]: https://stackoverflow.com/a/79055330/1420237

Author: lucaswerkmeister

src/HtmlParser.php

@@ -28,7 +28,7 @@ public function parseHtml( string $html ): DOMDocument {
 		// LIBXML_NOBLANKS Constant excludes "ghost nodes" to avoid violating
 		// vue's single root node constraint
 		if ( !$document->loadHTML( '<?xml encoding="utf-8" ?>' . $html, LIBXML_NOBLANKS ) ) {
-			//TODO Test failure
+			throw new Exception( 'Failed to parse HTML' );
 		}
 
 		/** @var LibXMLError[] $errors */
@@ -41,9 +41,15 @@ public function parseHtml( string $html ): DOMDocument {
 			libxml_disable_entity_loader( $entityLoaderDisabled );
 		}
 
+		$exception = null;
 		foreach ( $errors as $error ) {
-			//TODO html5 tags can fail parsing
-			//TODO Throw an exception
+			if ( strpos( $error->message, 'Tag template invalid' ) ) {
+				continue;
+			}
+			$exception = new Exception( $error->message, $error->code, $exception );
+		}
+		if ( $exception !== null ) {
+			throw $exception;
 		}
 
 		return $document;

tests/php/HtmlParserTest.php

@@ -63,4 +63,11 @@ public function testTwoRootNodes() {
 		$this->parseAndGetRootNode( '<p></p><p></p>' );
 	}
 
+	public function testMalformedHtml(): void {
+		$htmlParser = new HtmlParser();
+		$this->expectException( Exception::class );
+		$this->expectExceptionMessage( 'Unexpected end tag' );
+		$htmlParser->parseHtml( '</p>' );
+	}
+
 }