"Dimitry Sibiryakov" <SD@...> wrote:

>
> >Well, this makes some sense then. But doesn't it mean that users loose
> >any ability to work with charsets explicitly?
>
> No. We just have to map FB charsets to OS charsets.

Let's imagine we have to work with KOI-8R database on Windows (OS charset it
[directly] mapped to our WIN1251). We need proper sorting, so an
alphabetical collation is a requirement, hence we should setup the database
charset (or is it not necessary with unicode collations?). Data is stored
and passed to the client as UTF-8. Windows knows nothing about KOI8-R.
Should the client library translate UTF-8 to KOI8-R and then to WIN1251, or
directly UTF-8 to WIN1251?

And what about platforms compatibility? Does iconv exist in the default
installations of Solaris, HP-UX, AIX, FreeBSD and NetBSD? What about Darwin?


Dmitry