Since JRE 9 the Unicode Block "Cherokee Supplement" is supported by JRE. The Unicode Consortium mandates special case folding rules for "Cherokee" in the file CaseFolding-8.0.0.txt (see http://www.unicode.org/Public/8.0.0/ucd/CaseFolding.txt). Unfortunately, JRE 9 does not obey these rules, hence case folding works wrong in JRE 9.

Example: CaseFolding-8.0.0.txt mandates the outcome for case folding for code point 0x13F8 to be 0x13F0, but JRE 9 incorrectly produces 0x13F8 -- which simply is wrong.

Steps to reproduce:
* new String(new byte[] {(byte) 0x13, (byte) 0xF8}).toUpperCase().toLowerCase().codePoints();

Expected result:
* 13F0

Actual result:
* 13F8