Reputation: 121
I cannot figure out how to stop DOMDocument from mangling these characters.
<?php
$doc = new DOMDocument();
$doc->substituteEntities = false;
$doc->loadHTML('<p>¯\(°_o)/¯</p>');
print_r($doc->saveHTML());
?>
Expected Output: ¯(°_o)/¯
Actual Output: ¯(°_o)/¯
Upvotes: 12
Views: 6667
Reputation: 3252
PHP DOMDocument will not convert characters to htmlentities if the HTML is properly loaded in UTF-8 and has the meta charset=utf-8 tag.
The idea is to:
meta charset=utf-8
tag to the DOMDocumentmeta charset=utf-8
tag from after saving the result.Here's a sample code:
<?php
$htmlContent = file_get_contents('source.html');
$convertedContent = mb_convert_encoding($htmlContent, 'UTF-8', mb_detect_encoding($htmlContent));
$dom = new DOMDocument('1.0', 'UTF-8');
$dom->loadHTML($convertedContent);
// Create the meta tag element
$metaTag = $dom->createElement('meta');
$metaTag->setAttribute('http-equiv', 'Content-Type');
$metaTag->setAttribute('content', 'text/html; charset=utf-8');
// Append the meta charset tag to the head element
$head = $dom->getElementsByTagName('head')->item(0);
$head->appendChild($metaTag);
// Do any stuff here
// save the content without the meta charset tag
$new_content = str_replace('<meta http-equiv="Content-Type" content="text/html; charset=utf-8">', '', $dom->saveHTML());
// save to a destination file
file_put_contents('dest.html', $new_content);
Upvotes: 0
Reputation: 29932
I've found a hint in the comments of DOMDocument::loadHTML documentation:
(Comment from <mdmitry at gmail dot com> 21-Dec-2009 05:02: "You can also load HTML as UTF-8 using this simple hack:")
Just add '<?xml encoding="UTF-8">'
before the HTML-input:
$doc = new DOMDocument();
//$doc->substituteEntities = false;
$doc->loadHTML('<?xml encoding="UTF-8">' . '<p>¯\(°_o)/¯</p>');
print_r($doc->saveHTML());
Upvotes: 6
Reputation: 71
<?xml version="1.0" encoding="utf-8">
in the top of the document takes care of tags.. for both saveXML and saveHTML.
Upvotes: 3