Unicode upper/lower solution proposal

Hi all,

Disclaimer: I'm not a C programmer and my knowledge of locales is 
limited. I use Python.

PostgreSQL does not correctly use upper() and lower() for Unicode 
characters. I've read the bug reports and followups at 
archive.postgresql.org.

I'd like to propose two ideas:

1) Python has upper() and lower() functions, and they work in Unicode 
perfectly. With Python being open source, could PostgreSQL developers 
pick Python developer's brains? Is Python using only standard C library 
calls to do its Unicode upper/lower conversion? Could the relevant part 
of Python source code serve as a starting point of implementing Unicode 
upper/lower in Postgres?

2) Arbitrary collation support for PostgreSQL is available at 
http://www.fi.muni.cz/~adelton/l10n/ and it works perfectly. Could it be 
used to implement Unicode upper/lower in Postgres?

-- 
Milos Prudek