Tech Note: RSS Reader Rewritten
Technology | Fri, Nov 13, 2009 at 3:36:26 pm PST
Our left-column Ajax RSS widget, with a drop-down list of news sites and blogs that lets you check out their current RSS feeds, has been completely rewritten. The code now uses PHP’s native SimpleXML object library to parse all three flavors of RSS (0.92, 1.0, and 2.0) or ATOM feeds. Namespacing, XPATH, and attributes, oh my!
Another improvement fixes a small but annoying issue where strange characters would show up in place of em dashes, curly quotes, and other types of punctuation. These odd and ugly characters were caused by a difference in character encoding between the RSS feed and the LGF page where it’s being displayed. Rewriting everything to use my own RSS parsing code instead of an external library (like SimplePIE or MagpieRSS) helped me figure out the real reason for the problem. Once I understood it, the solution was amazingly simple. Just one line of code using the iconv function to convert the characters from UTF-8 to ISO-8859-1 encoding:
$theString = iconv("UTF-8", "ISO-8859-1//TRANSLIT", $theString);
The server-side PHP script also needs to send a ‘Content-Type’ header to set the character encoding to ISO-8859-1, before sending the formatted RSS data back to the calling Javascript routine. This ensures that all of our lovingly parsed characters look presentable.
header("Content-Type: text/html; charset=iso-8859-1");
My RSS script caches the HTML output of each feed for 30 minutes, to avoid abusing anyone’s bandwidth. But if you’re really craving the hottest, freshest news, you can manually refresh the feed before that time limit by clicking the ‘Reload’ button next to the drop-down list.
