It only seems to do that when I ask it to do RC extraction - it gives me another 3 nets, one of which is called

(null)

. Anyone seen this?

Full R-C extraction is iffy at best for a hierarchical design. The original code was written assuming a flat layout, and I have done some work to allow it to work in a hierarchical context, but there's a lot of work to be done, and meanwhile it does do things like output "null" nets. The question of when a subcell gets an additional port: Whenever magic detects something connected to a net in a subcircuit that is not declared a port. Doing parasitic capacitance extraction will often find lots of coupling between parent cell and subcell, which may cause it to add pins so that the parasitic cap has something to connect to. If you extract without parasitics, you will only get extra pins if something is really electrically connected to a net in a subcircuit that is not already marked as a port.

Ok that makes sense, but then why would it need to add a port to the top-level sub circuit deceleration? Since that is the highest level of hierarchy during the extraction.

That I don't know---it should not make any ports on the top level except for labels declared as ports.

Ok then thats odd - that is only for the full RC extraction mind you. Could it be related to the fact that you said there's still a lot of work to be done there?

Yes, sounds like it could very well be a bug.

ok cool - i won't worry about it then - cheers, Tim