Web crawlers and non-ASCII characters in sitemap.xml

One of our sites has non-ASCII (non-english) characters in URLs:

http://example.com/kb/начало-работ开发者_如何学运维ы/оплата

I wonder how do web crawlers (particularly Googlebot) handle these situations? Do these URLs have to be encoded or otherwise processed?

---

I think it is best url-encoded. This is the standard.