I use TYPO-3 CMS and it stores cyrillic filenames transcoded to unicode at my server when I upload files via this cms. Then apache reads these files perfectly fine (with no cms, just directly - the file is linked from the site like http://www.mysite.com/pics/%D0%BF%D0%BE). After I moved to another server, I get the problem Apache refuses to understand these unicode encoded files exist. In error log I see something like pics/\xd1\x88\xd0\xb5\xd0\xb9\xd0\xba at the malfunctioning server and something like this pics/%D0%BF%D0%BE at the correctly working server. How do I make my new Apache work and understand non-standart chars like my old one does? Literally I need apache to pass the query from the browser exactly as it appears, to the file system. But in beetwen Apache manages to do this replacement to the like of \xd1\x88\xd0\xb5\xd0\xb9\xd0\xba In every case I use the same browser (Firefox) while it should not influence, because each time I query not the cyrillic filename for the browser yet to transcode the way it wants, but already transcoded by the CMS, so every browser does the same query. One Apache understands the query, another does not.