peter_jacobi.rm wrote:

> Fortunately, charset and collation support
> someone are only semi-internal, as they are delegated
> to a DLL, so it's an easy start at Firebird hacking here.

I'll see into it as soon as I can spare some time. I should mention
though that I'll be working under linux and will thus be able to
generate only appropriate .so libs; I don't have the necessary tools for
the windows platform.

> Ho, hum! Challenge accepted. Do have a link? I'm generally good
> at googling around, but still have to find something concrete about
> collations.

Errr... Since the future seems to be unicode and Firebird does
transliteration through unicode mappings anyway, I went to see how
unicode guys handle collations. There seems to be just one collation
chart per character set, looks quite simple:
http://www.unicode.org/charts/collation/
Otherwise you might try http://www.alphabets-world.com/ or go to an
online directory (e.g. http://directory.google.com ), find some
government or linguistic pages of the country for which you seek
collation and try to find links to anything about its language. Usually
the alphabet should be sufficient to decipher the collation.

> So give me a link about Croation sort order if you find one.

http://www.hr/hrvatska/language/abeceda.en.htm
You might want to compare it with Slovenian alphabet/sort order:
http://www.ijs.si/slo-chset.html

-AlesS