> Can you explain what's the problem with the current configuration? Is
> it too hard to use a shared library or DLL that comes with the
> installation? Or the problem is that you don't want to do the UDF
> declarations in each database? I don't see any advantage in putting
> ib_udf inside the engine.

For example, JDBC specification defines standard functions in escaped
form. I can translate an escaped function into a it's corresponding
function in ib_udf, but I can do this only when user had declared them
in the database, otherwise I should throw an error. Should I for each
translation do a select on system table to check whether it is there?
Or should I disable all standard functions at all?

Claudio, there's not many functions in ib_udf to not inlcude it in the
engine (sooner or later). Why the hell do I need to declare abs() or
sqrt() in each database I need?

Roman