On Sat, Oct 17, 2009 at 2:57 AM, Andrew Dunbar hippytrail@gmail.com wrote:
I think it's a case of modern browsers behaving differently to older browsers. Older browsers only supported encoded URLs with % as the URL/URI standard is defined. But these are very user-unfriendly so modern browsers now convert these URLs into something readable for people whose native language does not use Latin script. I think Facebook only accepts URLs which comply to the standard and not the userfriendly human readable ones supported by modern browsers.
So it's no bug in Mediawiki and not really a bug in Facebook but it would be a userfriendly improvement for Facebook to interpret non-Latin URLs just as the modern browsers do.
URLs containing non-Latin characters are standardized, and known as IRIs. Lack of good IRI support in today's Web should probably count as a bug.