Bug Report: Non-ASCII Links Break
Use any link that uses characters not found in ASCII but found in Unicode in the forums and the link will truncate right before the first non-ASCII character. This is shown at http://www.duolingo.com/comment/1418492, where a link to http://en.wikipedia.org/wiki/Téléfrançais is not processed correctly. Because many links I have seen need to use Non-ASCII characters, this is an issue which needs to be fixed with the parser.
If you visit your site with Chrome (and possibly other browsers, but certainly not Safari) then copy the URL from the address bar it will be converted to an ASCII friendly format like this: http://en.wikipedia.org/wiki/T%C3%A9l%C3%A9fran%C3%A7ais.
Even though this is true, it is still a glitch in the link parser. Because of more non-latin alphabet and accented contents, this glitch in the parser should be fixed to meet up with current standards, because these are valid links.
Given the workaround I mentioned the only real problem I face is the treatment of the "&" character in URLs (this is really annoying because many sites with parameterised URLs use "&" as a separator). Anyway, sorry my comment wasn't useful.