FWIW: I wrote to the SQLite mailing list. And either I didn't describe the problem well enough, or there's no clear opinion about whether this is supposed to be working the way we expect it to work or not...
Hm... the conclusion basically is: localization is difficult. From the SQLite mailing list:
"Remember that fancy collations don't just look at 1 character at a time, they look at the whole thing, and can do surprising stuff based on that. In this case the order of preference for the collation looks like "when it's part of a larger word, then treating 'S' and 'Š' the same is more important than separating them. But when it's just 1 character then they're different. "
TBH: I have no idea how to fix/work around this right now. I'm sure it could be done, but I fear that performance would suffer heavily if we had to implement this locally in LMS.
-- Michael _______________________________________________ Squeezecenter mailing list [email protected] http://lists.slimdevices.com/mailman/listinfo/squeezecenter
