4D OCI and non Roman languages like Chinese or Russian

Product :4D for OCI
4D : v17.2 . 32bit
OS : Mac OS X 10.12
Oracle client: 11.2 Mac 32bit


I’m migrating a dictionary app from Sybase to Oracle database.
For the Sybase database connection we used the 4D LSQL plugging.
Now for the Oracle connection we are trying to use 4D for OCI.
The dictionary application is multilinguage so we need to retrieve records form the database in several Roman and non-Roman languages.
With 4D for OCI, at the begin, all accented characters in French, German, … were being replaced by the simple Ascii character and for languages like Russian or Chinese I was getting only rubbish.
After setting the NLS_LANG system variable to FRENCH_FRANCE.WE8ISO8859P1 I am now able to retrieve all the accented characters for the Roman languages, like French or German, but for the non-Roman I only get sequences of ‘¿¿¿’.
The Oracle database has been created with the UTF8 character set and I can see all the characters for all languages with SQL Developer but can’t find the good setting for the NLS_LANG system variable to have all characters visible in 4D. I tried with FRENCH_FRANCE.AL32UTF8 and the result was also bad.
Does anyone know which are the correct settings to have the Unicode UTF8 characters from the database to pass to 4D without any conversion?

Joaquim Franco