As has been noted before, EE’s search becomes case sensitive when it includes high-ascii characters. According to Rick:

MySQL knows how to map high-ascii values so that searches can be case-insensitive.  This isn’t really a limitation of EE, but rather of MySQL’s character support.

... it’s all MySQLs fault. OK, so this is why I am making it a feature request.

Feature Request: find a way to overcome this flaw, and provide truly case insensitive searches for foreign characters also. Perhaps a function could check whether a search string contains high-ascii characters, and either convert them to lowercase, or perhaps perform the search with all variations (ie lower and upper case umlauts), or whatever. I know too little of the internal workings of the search, but this has been bugging me for some time.

I second that

Yes, I too very much hope this will be solved some how in EE.

Been doing a bit of research and the solution is far cleaner if this is done on the MySQL server and instead of PHP.  Here is a nice thread asking about it with a nice response.

Still going to look into a PHP/EE solution but I think this is your best bet.

Another possibility is that if you are using MySQL 4.1 or above that you can go into your database tables and change your fields to have a collation that is for your particular language.  By default, MySQL uses the collation latin1_swedish_ci.  I switched that collation to latin1_german_ci, and the upper and lower case umlauts were matched at the same time, so success there.

Unfortunately not everybody has the possibility to recompile MySQL with custom swiches. But the second suggestion sounds helpful. I have noticed latin1_swedish_ci before, any idea why? I mean, why Swedish?

Quite a pity, I admit.  Man, if I knew half of the reasons why the MySQL blokes do what they do in new versions I would consider myself an Oracle of MySQL.

two years later, do we have some response from developers?

or we need to change collation? I’m puting that for last of last option.

d