Re: table functions

> FROM/WHERE: No, and I think it would be quite inappropriate for the
> function's behavior to depend on any such info.  To work correctly,
> you'd need to encode virtually a complete understanding of SQL into
> your

very true.  I could probably make a case for this being useful in
highly specialized functions.  But in general, I agree, not a good idea.

I've been working on a dblink clone that speaks JDBC via JNI, and am
just trying to figure out what postgres will (and won't) let me do.

> You can throw a
>     CHECK_FOR_INTERRUPTS();
> into your main loop whereever it seems safe to be killed by an
> interrupt.

cool!  Is this documented somewhere?  Did I overlook it?

thanks!

eric