DirectMemoryIO#getString(long, int, java.nio.charset.Charset) is broken for multi-byte encodings

[pepijnve]

The implementation of this method uses com.kenai.jffi.MemoryIO#getZeroTerminatedByteArray(long, int) to get a byte array containing the string data. For various encodings, like UTF-16, this does not work correctly. The code actually needs to look for the representation of the 0 code point in the encoding in question instead of just looking for a single 0 valued byte.

[pepijnve]

While digging through the code this morning I was looking at jnr.ffi.provider.converters.StringResultConverter#fromNative. This implementation does do the right thing and takes terminator width into account.

[pepijnve]

Duplicate of #30