DOM parser: remove certain attributes only

问题:

How can I use DOM parser to remove all attributes in span tags but except these two attributes,

<span style="text-decoration: underline;">cultura</span> accept

<span style="text-decoration: line-through;">heart</span> accept

reject this,

<span style="font-family: " lang="EN-US">May</span> accept

Is it possible?

My working code from the other post I made,

$content = '
<span style="text-decoration: underline;">cultura</span>l <span style="text-decoration: line-through;">heart</span>
<span style="font-family: " lang="EN-US">May</span>
';

$dom = new DOMDocument();
$dom->loadHTML($content);

foreach( $dom->getElementsByTagName( "span" ) as $span )
{

    foreach( $span->attributes as $attrib )
    {
        $span->removeAttributeNode( $attrib );
    }


}

$content =  $dom->saveHTML();

But this code will remove all attributes inside the span tags...

回答1:

You need to do it manually.

DOM handles HTML attributes, not CSS properties.

You need to access the style attribute, explode it's value using ; as a delimiter, then loop the array looking for the value you want to unset.

回答2:

This is entirely possible with DOM only. DOM exposes a parsed version of the style attribute through the element.style property.

$('*').each(function() {
    var s = this.style;
    for (var i = s.length - 1; i >= 0; i--)
       if (s[i] != 'font-weight')
           s.removeProperty(s[i]);
});

The above code removes every style except font-weight.

This uses jQuery's $('*') to iterate over every element, but of course you can do it the hard way using just DOM.

Here's a JsFiddle to play with: http://jsfiddle.net/NbN3S/

One difficulty with this is that the names appear to be browser-dependent, unfortunately: for example, text-decoration is taken apart by Firefox into a bunch of -moz-* styles.