There was a strange phenomenon this week at work when trying to access a certain PDF element from an AWS S3 bucket containing a German umlaut (ü) in the filename.
While on Chrome (50) and Firefox (45) it worked correctly (an XML answer showed correctly up, as it was expected in this case), it never worked on Internet Explorer (11).
After using the debug/developer tools (F12 in IE) I saw the following response which wasn't displayed in the browser itself:
InvalidArgumentHeader value cannot be represented using ISO-8859-1.response-content-dispositionattachment; filename=Eat_My_Zn?ni.pdf3E7CFF47D4A6BF23IfEECJc5hXFcmZXshxPKKJg/GCw+1gaWiyc+wXvvlmNTCJ7t3A8YrKARXI0tLnK0QkvAKnuNJBE=
So Internet Explorer 11 seems to have problems to encode the umlaut from ISO-88591- into UTF-8 (although this setting is activated in the Internet Options -> Advanced -> International).
When writing ue instead of ü it worked.
Update May 3rd 2016:
My colleague just found out that this only happens in a recent version of IE11. On Internet Explorer 11 Version 11.0.9600.16521 it was working. Another non-working version is 11.0.9600.17278 and the IE version I tested with was 11.0.9600.18282.