FreeTDS Development Group

[Niklas Lampén <Niklas.Lampen AT eventusmedia.fi>]

> > 1> UPDATE Mod_Texts SET ContentData='foo Найти' WHERE ContentId=6824;
> > 2> SELECT ContentData FROM Mod_Texts WHERE ContentId=6824;
> > 3> GO
> > ContentData
> > foo ?????
> 
> Is that field even capable of storing Cyrillic letters?

Yes, it is. The field is NVARCHAR and if I use Microsoft SQL Server 
Management Studio, I can insert that string in it and fetch it correctly with 
FreeTDS.

> It's possible it's not an NTEXT field and the server isn't set to use a
> Cyrillic single-byte character set.

SQL Server's 'sp_server_info' says that COLLATION_SEQ is 'charset=iso_1 
collation=Finnish_Swedish_CI_AS', which looks bad for this situation. But 
isn't all data stored as UCS2 in MSSQL and therefore it should be able to 
store those Cyrillic letters? And yet again, I can store 'foo Найти' in the 
field using Management Studio.

Is there a way to force FreeTDS to think that MSSQL Server is not ISO_1 but 
UCS2 or UTF-8 or something else? I would be happy to try to change some lines 
of code and recompile if someone just told me which lines to change.


Niklas