comparison with high unicode values fails [op/string-less-than.xml] K2-StringLT-1

[mschoeb]

The builtin java compareTo doesnt handle odd characters with very high unicode encodings correctly. It treats them as two parts and only compares the first one. Fix would definitely be possible but I dont think its worth it for now, since it could be quite a performance hit for every string comparison

"" < "𑅰"
should be true but it is false