GBJ corrupts characters in strings with characters with codepoints over 127
Strings containing characters with codepoints > 127 can become corrupted when transferred from the GBJ client to the server or from server to GBJ client. Corruption varies according to version of GBJ.
As of GBJ 3.0:
For extended ASCII characters (codepoint 128 to 255):
* Server-to-GBJ-client: No problems.
* GBJ-client-to-Server: Extended ASCII characters converted into UTF-8 multi-character codes.
For 16-bit characters (codepoint 256-0xFFFF):
* GBJ-client-to-Server: No problems, result is a DoubleByteString.
* Server-to-GBJ-client: Characters are byte-swapped on little-endian platforms (Linux, Windows).
Last updated: 10/6/16