Retrieving Arabic data from VARCHAR field

I'm trying to retrieve Arabic text data from an Oracle database owned and maintained
by another company.

The folks at that company apparently access the database from a Windows client, and they say that they enter and retrieve Arabic data just fine.

When I asked further, they said to just install Windows Arabic language support.

But our application (client) runs under Unix (Solaris 10)

I have tried using SQL*Plus and the C++ client library functions ResultSet::getString() and ResultSet::getUString()

In all cases, the Arabic characters are replaced by question marks. (The getUString command returns the same thing as getString(), except that it returns a string of 16-bit integers containing two characters

--
question marks and spaces --  per integer.)

I don't have any problem with the other character fields, which contain
normal ASCII characters.

The field itself is VARCHAR, length 60.


Any suggestions as to what we need to do to get Arabic?
I can deal with UTF-8, or UCS-16, or even some Microsoft code page
if I have to.

I just can't convert a string of question marks into meaningful Arabic.


The database (server) is vers. 10.2.0.1.0
We are using the Solaris 32-bit  InstantClient libraries vers, 10.2.0.2

-- Alan McKenney