diff options
Diffstat (limited to 'src/HTMLParser.php')
-rw-r--r-- | src/HTMLParser.php | 18 |
1 files changed, 16 insertions, 2 deletions
diff --git a/src/HTMLParser.php b/src/HTMLParser.php index 21bb88f..5a684a5 100644 --- a/src/HTMLParser.php +++ b/src/HTMLParser.php @@ -438,7 +438,8 @@ class HTMLParser $append = false; // TODO Check if this comparison working as expected - if ($sibling === $topCandidate) { + // On the original js project it was a simple $sibling == $topCandidate comparison. + if ($this->compareNodes($sibling, $topCandidate)) { $append = true; } else { $contentBonus = 0; @@ -473,7 +474,7 @@ class HTMLParser // $sibling->setNodeName('div'); } - $import = $articleContent->importNode($sibling->getDOMNode()); + $import = $articleContent->importNode($sibling->getDOMNode(), true); $articleContent->appendChild($import); } } @@ -504,4 +505,17 @@ class HTMLParser } return false; } + + private function compareNodes($node1, $node2) + { + if ($node1->getTagName() !== $node2->getTagName()) { + return false; + } + + if ($node1->getTextContent() !== $node2->getTextContent()) { + return false; + } + + return true; + } } |