"Dimitry Sibiryakov" <aafemt@...> wrote:

>
> Besides, there are things that Y-valve can count itself without
> bothering engines: it has full information about connections,
> transactions, queries and so on.

But not about databases or users. The info requests cannot be handled only
by the Y-valve. Also, AFAIK the Y-valve has no idea whether it runs on the
client side or on the server. But the client side itself cannot process your
request while the server one could. So the Y-valve *must* delegate the call
to some provider.

> But it should implement fb_engine_info() call which has no
> considerable differences from SAPI.

It's dumber, as it should not make any decisions itself.

> Do you mean that fb_engine_info() can be called from remote site?

It depends on how we design the things. I don't want to see the host name
passed to fb_engine_info() as it duplicates SAPI.


Dmitry