Alexey,
>I'm using extension for base unicode support
>(http://mobigroup.ru/files/sqlite-ext/unicode/), but in last two
>releases find the problem with indexes by columns with redefined
>NOCASE collation
This code has many problems and the version on your site (the same
version is available elsewhere) doesn't work as expected. NOCASE is
not the only function with problems, there are more with UPPER,
LOWER. Even some tries have wrong data. I had to recompile three of them.
I was in need of such extension (and a bit more), started looking at
open source code and then began work on it. Unfortunately I had to
switch to other boring but urgent tasks, and for really much longer
than I would have liked.
I now have some time to finish it and it should be ready for beta
anytime soon.
The module offers some universal Unicode support:
UPPER
LOWER
UNACCENT
FOLD
LIKE (unaccented version)
GLOB (unaccented version)
TYPOS (unaccented lowercased Damerau-Levenshtein distance on strings
with support for '_' and trailing '%' as in LIKE)
NOCASE
LETTERS (a simple wrapper to a Windows function for locale-independant
unaccented collation)
specific handling for small German sharp s 'ß'
I compile for Windows 32 with MinGW gcc but it shouldn't be very hard
for someone to make it work on another OS as well.
JcD
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
