Test and fix getRawDecodedValueFromHtml method

contao · May 19, 2021 · 5821b91 · 5821b91
1 parent f7969ba
commit 5821b91
Show file tree

Hide file tree

Showing 2 changed files with 57 additions and 2 deletions.
diff --git a/core-bundle/src/Resources/contao/library/Contao/StringUtil.php b/core-bundle/src/Resources/contao/library/Contao/StringUtil.php
@@ -1024,14 +1024,19 @@ public static function getRawDecodedValue(string $strValue, bool $blnRemoveInser
 	 */
 	public static function getRawDecodedValueFromHtml(string $strValue, bool $blnRemoveInsertTags = false): string
 	{
+		if (!$blnRemoveInsertTags)
+		{
+			$strValue = Controller::replaceInsertTags($strValue, false);
+		}
+
 		// Add new lines before and after block level elements
 		$strValue = preg_replace(
-			array('/[\r\n]+/', '/</?(?:br|blockquote|div|dl|figcaption|figure|footer|h\d|header|hr|li|p|pre|tr)\b/i'),
+			array('/[\r\n]+/', '/<\/?(?:br|blockquote|div|dl|figcaption|figure|footer|h\d|header|hr|li|p|pre|tr)\b/i'),
 			array(' ', "\n$0"),
 			$strValue
 		);
 
-		$strValue = static::getRawDecodedValue($strValue, $blnRemoveInsertTags);
+		$strValue = static::getRawDecodedValue($strValue, true);
 
 		// Remove duplicate line breaks and spaces
 		$strValue = trim(preg_replace(array('/[^\S\n]+/', '/\s*\n\s*/'), array(' ', "\n"), $strValue));

diff --git a/core-bundle/tests/Contao/StringUtilTest.php b/core-bundle/tests/Contao/StringUtilTest.php
@@ -12,11 +12,13 @@
 
 namespace Contao\CoreBundle\Tests\Contao;
 
+use Contao\CoreBundle\Security\Authentication\Token\TokenChecker;
 use Contao\CoreBundle\Tests\TestCase;
 use Contao\StringUtil;
 use Contao\System;
 use Psr\Log\NullLogger;
 use Symfony\Component\DependencyInjection\ContainerBuilder;
+use Symfony\Component\HttpFoundation\RequestStack;
 
 class StringUtilTest extends TestCase
 {
@@ -26,6 +28,10 @@ protected function setUp(): void
 
         $container = new ContainerBuilder();
         $container->setParameter('kernel.project_dir', $this->getFixturesDir());
+        $container->setParameter('kernel.cache_dir', $this->getFixturesDir().'/cache');
+        $container->setParameter('kernel.debug', false);
+        $container->set('request_stack', new RequestStack());
+        $container->set('contao.security.token_checker', $this->createMock(TokenChecker::class));
         $container->set('monolog.logger.contao', new NullLogger());
 
         System::setContainer($container);
@@ -142,4 +148,48 @@ public function trimsplitProvider(): \Generator
             ['foo', 'bar'],
         ];
     }
+
+    /**
+     * @dataProvider getRawDecodedValueProvider
+     */
+    public function testGetsRawDecodedValues(string $source, string $expected, bool $removeInsertTags = false): void
+    {
+        $this->assertSame($expected, StringUtil::getRawDecodedValue($source, $removeInsertTags));
+    }
+
+    public function getRawDecodedValueProvider(): \Generator
+    {
+        yield ['foobar', 'foobar'];
+        yield ['foo{{email::test@example.com}}bar', 'footest@example.combar'];
+        yield ['foo{{email::test@example.com}}bar', 'foobar', true];
+        yield ['{{date::...}}', '...'];
+        yield ['{{date::...}}', '', true];
+        yield ["&lt;&#62;&\u{A0}[lt][gt][&][nbsp]", "<>&\u{A0}<>&\u{A0}", true];
+        yield ["I &lt;3 Contao", "I <3 Contao"];
+        yield ["Remove unexpected <span>HTML tags", "Remove unexpected HTML tags"];
+        yield ["Keep non-HTML &lt;tags&#62; intact", "Keep non-HTML <tags> intact"];
+        yield ["Cont\xE4o invalid UTF-8", "Cont\u{FFFD}o invalid UTF-8"];
+        yield ["&#123;&#123;date&#125;&#125;", "\u{200C}{\u{200C}{date}}"];
+    }
+
+    /**
+     * @dataProvider getRawDecodedValueFromHtmlProvider
+     */
+    public function testGetsRawDecodedValuesFromHtml(string $source, string $expected, bool $removeInsertTags = false): void
+    {
+        $this->assertSame($expected, StringUtil::getRawDecodedValueFromHtml($source, $removeInsertTags));
+    }
+
+    public function getRawDecodedValueFromHtmlProvider(): \Generator
+    {
+        yield from $this->getRawDecodedValueProvider();
+
+        yield ['foo<br>bar{{br}}baz', "foo\nbar\nbaz"];
+        yield [" \t\r\nfoo \t\r\n \r\n\t bar \t\r\n", "foo bar"];
+        yield [" \t\r\n<br>foo \t<br>\r\n \r\n\t<br> bar <br>\t\r\n", "foo\nbar"];
+        yield [
+            "<h1>Headline</h1>Text<ul><li>List 1</li><li>List 2</li></ul><p>Inline<span>text</span> and <a>link</a></p>",
+            "Headline\nText\nList 1\nList 2\nInlinetext and link",
+        ];
+    }
 }