Fixed decoding img src attributes.

parpalak · Feb 11, 2024 · c4ee100 · c4ee100
1 parent 275915c
commit c4ee100
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/src/S2/Rose/Extractor/HtmlDom/DomExtractor.php b/src/S2/Rose/Extractor/HtmlDom/DomExtractor.php
@@ -246,7 +246,7 @@ protected static function processDomElement(\DOMNode $domNode, DomState $domStat
 
             case 'img':
                 $domState->attachImg(
-                    html_entity_decode($domNode->getAttribute('src'), ENT_QUOTES | ENT_SUBSTITUTE | ENT_HTML5),
+                    rawurldecode(html_entity_decode($domNode->getAttribute('src'), ENT_QUOTES | ENT_SUBSTITUTE | ENT_HTML5)),
                     $domNode->getAttribute('width'),
                     $domNode->getAttribute('height'),
                     html_entity_decode($domNode->getAttribute('alt'), ENT_QUOTES | ENT_SUBSTITUTE | ENT_HTML5)

diff --git a/tests/unit/Rose/IntegrationTest.php b/tests/unit/Rose/IntegrationTest.php
@@ -260,7 +260,7 @@ public function testFeatures(
 
         $img1 = $resultSet11->getItems()[0]->getImageCollection()->offsetGet(1);
         $this->assertNotNull($img1);
-        $this->assertEquals('2.jpg', $img1->getSrc());
+        $this->assertEquals('2 3.jpg', $img1->getSrc());
         $this->assertEquals('20', $img1->getWidth());
         $this->assertEquals('25', $img1->getHeight());
         $this->assertEquals('Alternative text', $img1->getAlt());
@@ -424,7 +424,7 @@ public function indexableProvider()
             ,
             (new Indexable('id_1', 'Another instance', 'The same id but another instance. Word "content" is present here. Twice: content. Delimiters must be $...$ or  \[...\]', 20))
             ,
-            (new Indexable('id_4', 'Another instance', 'Nothing is here but images: <img src="1.jpg" width="10" height="15"> <img src="2.jpg" width="20" height="25" alt="Alternative text" />', 20))
+            (new Indexable('id_4', 'Another instance', 'Nothing is here but images: <img src="1.jpg" width="10" height="15"> <img src="2%203.jpg" width="20" height="25" alt="Alternative text" />', 20))
             ,
         ];