DoubleByteStrings cause server compiler errors

The way DoubleByteStrings are handled by the GCI changed in 3.0; as a result,
TwoByteStrings from GBS little-endian clients (Windows and Linux) are incorrectly
swizzled, resulting in server compiler errors when attempting to compile
source code that is a DoubleByteString.

Since any use of a Character outside the single byte range modifies the
underlying VW tools to use a TwoByteString, you will also see compiler
errors performing any evaluate from a workspace whose text previously included
a large Character (for example, the result of a print-it). This error
occurs even though the code being evaluated has no large characters.

Workaround:

none/not applicable