>> In test 4you compare lower letter "f" with upper letter "F" then you got
>> difference.
>> All is ok in this test case

> The collation being used should be case insensitive. Test 4 results show it
> is not, so all is not well.

> Or am I reading this wrong?

No. Test 3 proves that the case insensitive comparison for 'Fu' and
'fu' works correctly. But as soon as you include the "Sharp S"
character (comparing 'Fuß' with 'fuß', Test 4) it doesn't work.

There is no such thing as an uppercase Sharp S *) so maybe the lookup
tables for case insensitiveness or accent insensitiveness are wrong in
that place?

Regards

Stefan


*) The German "Sharp S" is derived from a ligature between a Fraktur
small "long s" and a small z (that's wy it's also called "Eszet" in
German). Don't be fooled by the Unicode Consortium having defined a
"U+1E9E LATIN CAPITAL LETTER SHARP S" - this is a completely abstract
concept, there is no such letter (or glyph). In cases where only
uppercase letters are available, the Sharp S is replaced by an "SS"
sequence (or, where this could be misleading or ambuguous, by an "SZ"
sequence). In Switzerland, the Sharp S is not used at all.