Currently, QuickEvent is fine-tuned for Latin-based scripts. Characters are stripped of diacritics and narrowed down to 7-bit ASCII. This makes filtering, searching and other string operations typo-tolerant. But the collation algorithm fails dramatically with any other script, for instance, Cyrillic. When I try to filter names, a match is found in every row.
I ended up with a temporary fix to just compare strings ordinarily, and this works just fine: 6177c5e.
But a more robust solution acceptable for anyone could be a choice of either the Latin-based collation or the generic one. This could be a configuration option or automatic selection based on the current locale, maybe.
Currently, QuickEvent is fine-tuned for Latin-based scripts. Characters are stripped of diacritics and narrowed down to 7-bit ASCII. This makes filtering, searching and other string operations typo-tolerant. But the collation algorithm fails dramatically with any other script, for instance, Cyrillic. When I try to filter names, a match is found in every row. I ended up with a temporary fix to just compare strings ordinarily, and this works just fine: 6177c5e. But a more robust solution acceptable for anyone could be a choice of either the Latin-based collation or the generic one. This could be a configuration option or automatic selection based on the current locale, maybe.