Weird behavior with Unicode data

Hello everyone.

I have detected a rather weird behavior when I use FreeTDS with PHP on a
CentOS machine and perform a select on SQL Server 2012 on Unicode data. I
have nvarchar(50) columns that contain Unicode strings with variable
lengths. When any string has a length of more than 25, the driver does not
return the column data at all, returning an empty string instead.

I am using the latest FreeTDS driver with the latest unixODBC and I have
set the tds version to 8.0 and the client charset to utf-8. When the client
charset is not set the response is a lot of ??? which is probably logical.

Is there anything that anyone has in mind? I would happily provide any
additional info you might need.

Thanks in advance for any help that you might be able to provide.


Savvas Kampouridis