Errors on UTF8 conversion on Linux, Solaris/Intel
Attempts to read/write UTF8 Oracle data on Linux and Solaris/Intel platforms
do not correctly byte-swizzle the characters. This can result in incorrect
results or #typeConversionErrors.
In the case of CLOBs, if *only* GemConnect is used to access the Oracle
data, the results in GemConnect will be correct, although the data will
be stored incorrectly in Oracle. But incorrect results will occur if other
external processes read/write the Oracle data.