Bizarre "@namespace" and styling information in XML document

Question

I am getting an XML document like this:

<?xml version="1.0" encoding="UTF-8"?>
@namespace html url(http://www.w3.org/1999/xhtml); :root { font:small Verdana; font-wei.... huge list of styling
<items>
    <item>
    ...

That second line seems to be preventing me from parsing the file.

Using Tidy

<?php
$config = array(
       'indent'     => true,
       'input-xml'  => true,
       'output-xml' => true,
       'wrap'       => false);
$tidy = new tidy;
$tidy->parseFile('https://website.com/path/to/XML.ashx?param=12345', $config);
$tidy->cleanRepair();

print_r($tidy);
?>

which will result in:

tidy Object
(
    [errorBuffer] => 
    [value] => <?xml version="1.0" encoding="utf-8"?>

)

Using simplexml_load_file()

<?php
$xml = simplexml_load_file('https://website.com/path/to/XML.ashx?param=12345');
print_r($xml);
?>

output:

**Warning**: simplexml_load_file(): https://website.com/path/to/XML.ashx?param=12345:1: parser error : Start tag expected, '<' not found in **C:\xampp\htdocs\local\php\script.php** on line 2

**Warning**: simplexml_load_file(): <?xml version="1.0" encoding="utf-8" ?> in **C:\xampp\htdocs\local\php\script.php** on line 2

**Warning**: simplexml_load_file(): ^ in **C:\xampp\htdocs\local\php\script.php** on line 2

I've also tried various cURL options and simply file_get_contents()

My question is: What is that second line of XML and how can I parse this file?

Answer 1

XML does not allow non-whitespace textnodes after the XML-Declaration. So what you have is invalid XML and this it what the libraries are telling you. But Tidy (release 25 March 2009) can deal with that:

$buffer = '<?xml version="1.0" encoding="UTF-8"?>
@namespace html url(http://www.w3.org/1999/xhtml); :root { font:small Verdana; font-wei.... huge list of styling
<items>
    <item></item> </items>';

$config = array(
    'indent'     => true,
    'input-xml'  => true,
    'output-xml' => true,
    'wrap'       => false);
$tidy = new tidy;
$tidy->parseString($buffer, $config);
$tidy->cleanRepair();

print_r($tidy);

Output:

tidy Object
(
    [errorBuffer] => line 2 column 1 - Warning: discarding unexpected plain text
    [value] => <?xml version="1.0" encoding="utf-8"?>
<items>
  <item></item>
</items>
)

So you most likely have more issues with that "XML" (or it's a limitation of a buffer if you have a very very large line there).

As this isn't XML you might ask yourself what that is? It's CSS, and what you have there is a so called at-rule^Q&A, more concrete a CSS Namespace Declaration. (Browsers (user-agents) per the earlier CSS specs did not have to support any of these. Even the current CSS Selector API requires any namespace prefix resolutions to cause an exception within the API. A good example of CSS namespace usage with a XML (XHTML) document is in this earlier answer).

What follows in your chunk of text is a namespace prefix and the CSS under it.

So what you have there is a mixture of different data. It won't parse valid as XML and you won't find any common browser that actually can deal with that CSS either - even if it would validate - because it's not clear that that text is CSS (it would need to be wrapped inside an element denoting a stylesheet).

Side-Note: A correct CSS parser would drop the XML here as it's invalid and the CSS specs denote that anything invalid needs to be dropped. So what you have there - in the whole - could technically conform as CSS document. You think it's XML, it's just CSS ;)

So as bizarre as this @ rule might sound to you, it actually isn't. It exists, just not at such a place.

On the other hand it's not really helpful to cover up the source as website.com - seeing the real site might have given more context to tell you more.

Answer 2

It is not XML it is XHTML (Extensible HyperText Markup Language) and the below is declaring the html namespace, followed by CSS styling. Hence your browser understands it but an XML parser won't.

@namespace html url(w3.org/1999/xhtml); 

It is HTML meant to be compliant with XML however it looks like this page may not compliant to the strict XHTML and hence not parsing as XML.