Strange character in XML document

I have a strange character showing up on my rss fee开发者_JS百科d. On firefox, it looks like a box with four numbers in each corner, in some cases, 0 - 0 - 9 - 4, in other 0 - 0 - 9 - 2.

These are appearing where smart quotes should be.

I'm familiar with the black diamond with the question mark, but this is a new one.

---

The 0-0-9-4 indicates that the character was 0x0094, which is an unassigned UTF-16 character. Whatever is producing the feed is inserting characters for which your browser has no font mapping, or possibly the character-encoding specified in the header doesn't match the stream contents.

---

Ah, okay. You pointed my in the right direction. What was coming up was Windows entities. People put stuff into our database in a complex series of steps converting from Word, to InDesign, to GoLive (yes, it is painful).

Anyway, what the database was popping out was these entities like '’', which mean something I guess to windows, but nothing to my browser, in either ISO-8859-1 or UTF-8, so no amount of changing my page encoding could fix that nonsense. Though, oddly, it just appeared here correctly, so I don't know what I'm doing wrong.

So anyway, I fixed it by running everything through this php function before displaying it.

function fixChars($text){

    // Next, replace their Windows-1252 equivalents.
    $text = str_replace(
    array('‘', '’', '“', '”', '•', '—', '…'),
    array("'", "'", '"', '"', '-', '--', '...'),
    $text); 

    return $text;

}

So, now things seem fine.

Thanks for the direction all.