At 03:49 PM 23/09/2004 +0000, you wrote:

>The record's content "Kovács Ágnes" uppered by FB, it maked to:
>KOVáCS ÁGNES (because á,Á is not in range A-Z,a-z.)
>
>But my select parameters are uppered as:
>KOVáCS áGNES.
>
>And KOVáCS ÁGNES<>KOVáCS áGNES.
>
>The "uppering" is not modifiable by "collate" ?

These problems with PXW_HUN are probably the reason Sandor Szollosi made
the ISO_HUN collation for character set ISO8859_2. You should probably
begin again and declare ISO8859_2 as your charset. btw, ISO_HUN is not (as
described in the earlier release notes) a case-insensitive collation. It
is a Hungarian dictionary collation.

> >
> > >(I need to working with Linux and Windows, so if I create UDF-s, that
> > >everytime I need to redeclare UDF-s, when I copy the DataBase
>between OSs)
> >
> > ????
>
>So: when I write an UDF in Windows by Delphi, it is use DLL. When I
>write an UDF in Linux, it is use SO.
>I working in localhost (Win), and when I finished with module, I
>published in Linux.
>But in Linux many user editing the records.
>When I want to copy the DB to my machine, to work with fresh datas,
>the UDF-s are points wrong path (to an SO).
>So I must redefine the UDF's SQL interface by hand after ds file copy.

No, you are doing that wrongly. It is *your* choice to include the file
extension but, as you have discovered, you spoil cross-platformness by
doing that. Leave it out and save yourself all that trouble!

./heLen