My understanding was that the "international" branch fixed
implementation bugs that impacted all multi-byte character sets,
including unicode_fss. Those corrections are included in FB2. Unless I
missed a patch (entirely possible), those corrections are not included
in any standard FB1.x build.

As you pointed out in another post, FB2 is beta and should not be
considered stable. But it is really sweet!

On Mon, 2006-05-22 at 23:57 +1000, Helen Borrie wrote:
> Contrary to what Adam stated, the UNICODE_FSS implementation is not
> "incomplete / doesn't work properly". It works as designed. It's
> just that it is a very limited implementation - all characters are a
> static 3 bytes and there is no upper/lower case mapping. But if the
> client is correct then it should return the unicode_fss characters as
> stored.
>