So about this one, I don't feel like doing this change now as the values of `areStringComparisonsCaseInsensitive()` doesn't seem very reliably set. Especially, it's always true for MySQL whereas it's not the case at all and depends on the collation.

I expect we would break a lot of applications.

It might be interesting to revisit this for 6 though if we find a reliable way to get the information from the database.