I have a database that I have been using with an old  Borland C++ Builder 4 project, using IBObjects 4.3Aa.

The database was created with character set "NONE".  In my project, the TIBODataset has CharSet left as undefined (null).

Everything has been working fine in the application itself for years.

Now, I'm writing a .NET application using the Firebird .NET provider that talks to the same database.  Unfortunately, it is messing up my data when there are special characters involved.  If it reads a record with a special character (like the trademark symbol), then writes it back to the database, those special characters are getting written back as some other character (usually multiple characters), and the look like garbage when I read them back in my IBO application.

So, it looks like an issue of mismatched character sets.  In my .NET application, in the connection string, I have the character set defined as "NONE", but I still have the above error.  I tried setting it to UTF8 in the .NET applicaiton, and I still have the same issue (but with different characters being written).

I guess I'm not real clear what a character set of "NONE"/null means.  

I guess my question is: What character set gets used if I have the database with the character set defined as "NONE" and IBO set to null?

I can't really easily change the database or the IBO application at this point to specify a character set, as we have many many users with many many years of data stored the way it is now.  I need to find out what character set I can set the .NET application to so that the characters will be compatible with my IBO project.

Thanks,

   Joe