PHP DOM extension, "loadHtml()" problem

vincent_L

Hi everyone !

I try to get some Web pages, and parse it with the PHP DOM extension.

My problem is that PHP-DOM seems to be confuse with anchors.

For example, get a Wikipedia page (http://fr.wikipedia.org/wiki/Linux ) and load it with DomDocument :

$snoopy = new Snoopy;
$snoopy->fetch("http://fr.wikipedia.org/wiki/Linux");

$htmltext = $snoopy->results;
$domdoc = DOMDocument::loadHtml($htmltext);

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: ID top already defined in Entity

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: ID Histoire already defined in Entity

Warning: DOMDocument::loadHTML() [function.DOMDocument-loadHTML]: ID Le_projet_GNU already defined in Entity

etc...

All these attributes are anchors : <a href="#Histoire"> , <a href="#Le_projet_GNU">, etc.

I tried to change the encoding system and some other crap-tricks ;-) but these messages are still there.

Anybody has suggestion ?

MarkR

I BELIEVE this is because the elements that these anchors point to have both name and ID set to the same thing, and libxml2 attempts to treat the name additionally as an ID, resulting in a duplicate ID. It's not because of the anchors:

<p><a name="Histoire" id="Histoire"></a></p>

Which may generate this warning.

Normally, I disable warnings when running loadHTML, as most pages generate a lot of them.

Mark

vincent_L

First, thanks for your reply.

MarkR wrote:
Normally, I disable warnings when running loadHTML, as most pages generate a lot of them.

You mean, using validateOnParse ?

$domdoc = new DOMDocument;
[B]$domdoc->validateOnParse = false;[/B]
$domdoc->loadHtml($htmltext);

Humm....but some functions needs the document to be validated, like getElementById, am I wrong ?

MarkR

No, I normally disable warnings by using error_reporting, then reset error_reporting after I've done loadHTML.

Mark