As Martin pointed out, the UTF-16 variants probably are not critical
enough to have "special"
fastpath implementation in StringCoding. If better performance of these
UTF-16 charsets is
really desired, it might be worth implementing the
sun.nio.ArrayDe/Encoder interface in
UnicodeDe/Encoder to
Very high level:
UTF-16 is not expected to be a popular encoding for text outside the JDK.
Everyone is supposed to be migrating to UTF-8 from ISO-8859-1 and other
legacy encodings.
The fact that people (like you and I) are writing specialized
encoders/decoders outside of the "real" charset implem
I was looking at the OpenJDK 9 code, and I noticed that optimizations for
encoding and decoding from UTF-16 text could be added to the
java.lang.StringCoding class.
Here is how the optimized UTF-16 decoding could be implemented in
java.lang.StringCoding:
private static void byteSwapUTF16(byte[]
