I was doing some web programming for Windows Mobile device, and was facing this problem on Pocket Internet Explorer.

The problem is, for all modern browser, URL request will be encoded and sent in UTF-8, Oprea for PPC (8.65), Netfront (4.5TP) does handle that correctly, only PIE fails to do that, here is the example:

when you process URL:
http://www.google.com.hk/search?q=中文
Browser should encode the URL and process:
http://www.google.com.hk/search?q=%A4%A4%A4%E5

But in PIE it does not handle it at all, resulting in error on the request.

Many developers handle non-ASCII chars in URL by using escape in JavaScript, in the case PIE fail to handle it too, instead of RFC 2279 standard, %nn%nn, it process the escape as %unnnn which was the very old and outdated way of how it should be done. (I consider something that was already fixed in MSIE 5.x and Netscape 4.x was outdated.)

Instead of using escape(foo), use encodeURIComponent(foo) before using the string as URL fix the problem.

Links:
UTF-8 Encoding for Older Browsers – WorldTimeZone