Fixed Invalid characters being allowed in page nodes

[Stuart Wright]

Error

The provided file was not a valid XML file. Please provide a valid XML file.

I got this error when exporting our live style and importing it to overwrite a sibling style.
I ran an XML validator on the XML and it threw up an error:
Errors in the XML document:
13429:66An invalid XML character (Unicode: 0xb) was found in the CDATA section.


The source from this page was copied from a PDF into Dreamweaver, formatted and then pasted into the page node.
Now this page has been giving me trouble for a few days. Basically any changes I make in the page node edit tab aren't being updated. The usual 'working' animation happens and the slide-down notice that the page has been saved. But it stays the same when I refresh it.

http://www.avforums.com/pages/advertise/

The only way I can actually update the page is by editing the template _page_node.504.

So first of all there is a bug, I believe, which is allowing crap characters to be saved in page nodes which invalidate the XML.
But also there is something wrong with this page which means I can't save it. The one issue might be related to the other. This is the only page I am having problems with.

 

[HWS]

Have you checked your XML to be UTF-8 formated?

 

[Stuart Wright]

Have you checked your XML to be UTF-8 formated?

This is the XML produced by the Xenforo export style function.

 

[Mike]

Can you paste those characters here? Looking at the HTML, there are no odd characters where the data is highlighted (just plain ASCII).

 

[Stuart Wright]

I removed it from the page Mike in an attempt to get it to save successfully (it didn't).
I've attached the text which I pasted into Dreamweaver and then the page node. The odd character is before the word 'each' on the penultimate line.

 

[Mike]

I've now changed our string clean up functions (on input) to deal with some additional control characters. This won't be retroactive though (the character will be removed next time the browser sends it to the server).