Ok folks, I'm at the end of my tether here.

I've tried using IB6, but it does not have
a Hungarian dictionary sort collation.

I've tried using FB1.02 which *_says_* it has
a Hungarian dictionary collation, but it is
not in fact a Hungarian dictionary collation.

Now, I'm trying FB1.52 (impressive reading of
the docs, BTW!).

Using IBOConsole when I go to choose a CHARACTER_SET
for my db, I am not allowed choose ISO8859_2, which
supports a proper Hungarian collation (ISO_HUN), so
I create one with ISO8859_1.

Then, in this just created db, I check in
RDB$Character_Sets and what do I find, but
ISO8859_2, the very one I want. So, I try to
be smart, and create the db with a script this
time, instead of using the IBOConsole interface.

---------------------
SET SQL DIALECT 3;

CREATE DATABASE 'D:\Paul\Hungarian\Dictionary\Dictq.fdb'
-user 'SYSDBA' -password 'masterkey'

PAGE_SIZE 4096
DEFAULT CHARACTER SET ISO8859_2;
------------------------

Works a treat. So, I say to myself, the character
set list is somehow hardcoded into IBOConsole, but
I have managed to get around that. But, when I try
to create my table with the ISO_HUN collation, it
fails with the message

----------------------------
ISC ERROR CODE:335544569

ISC ERROR MESSAGE:
Dynamic SQL Error
SQL error code = -204
Data type unknown
COLLATION ISO_HUN is not valid for specified CHARACTER SET
-------------------------------

I know that you're all probably sick of me
at this stage, but is there *_any_* way,
(kludge, trick, voodoo, sacrificing of virgins)
that I can get ISO8859_2 with ISO_HUN to work?

It's driving me up the !"£$%^& wall, that I can't
get it to work, no matter what combination of IB, FB
IBConsole, IBOConsole (not to name the other tools
that I've tried).

I would also humbly suggest that it's a (big) bug
that a feature which Firebird has advertised as
being in 1.02 (which it isn't) is not in FB 1.5.2.4731
either.

Thanks for any info, pointers or anything that
might help!!!


Paul...


p.s. I am just a humble applications programmer,
so suggesting RTFS is not much of a help to me.
I could, perhaps, if sent the source for the
WIN1250 CHARACTER_SET with the PXW_HUNDC (the
supposed Hungarian dictionary sort) could try
and work out what's wrong with it. But recompiling
the entire server is *_way_* beyond my
capabilities.

In fact, I think I'll download the source now and
see if I can find the bit that I want, then
maybe I could send in modified code to be included
in the next new build - would this be acceptable?