[Crawler] Fix regression where cdata nodes will return empty string

symfony · Aug 1, 2023 · 3fdd2a3 · 3fdd2a3
1 parent 8aa333f
commit 3fdd2a3
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/Crawler.php b/Crawler.php
@@ -588,7 +588,7 @@ public function innerText(/* bool $normalizeWhitespace = true */): string
         $normalizeWhitespace = 1 <= \func_num_args() ? func_get_arg(0) : true;
 
         foreach ($this->getNode(0)->childNodes as $childNode) {
-            if (\XML_TEXT_NODE !== $childNode->nodeType) {
+            if (\XML_TEXT_NODE !== $childNode->nodeType && \XML_CDATA_SECTION_NODE !== $childNode->nodeType) {
                 continue;
             }
             if (!$normalizeWhitespace) {

diff --git a/Tests/AbstractCrawlerTestCase.php b/Tests/AbstractCrawlerTestCase.php
@@ -377,6 +377,12 @@ public static function provideInnerTextExamples()
                 '',
                 '  ',
             ],
+            [
+                '//*[@id="complex-elements"]/*[@class="six"]',
+                'console.log("Test JavaScript content");',
+                'console.log("Test JavaScript content");',
+                ' console.log("Test JavaScript content"); ',
+            ],
         ];
     }
 
@@ -1311,6 +1317,7 @@ public function createTestCrawler($uri = null)
                         <div class="three"> Parent text <span>Child text</span> Parent text </div>
                         <div class="four">  <span>Child text</span>  </div>
                         <div class="five"><span>Child text</span>  <span>Another child</span></div>
+                        <script class="six" type="text/javascript"> console.log("Test JavaScript content"); </script>
                     </div>
                 </body>
             </html>