RE: [sv-bc] .name and .*

I remember that there was discussion of this, during which I learned
that the intent was to require the name to exist in the instantiating
scope, for each port that was being connected by one of these.

And the LRM does effectively say that, when it says that this is
equivalent to the named connection but does not create an implicit
net.  As you say, "On the face of it, that should cause an error."
And so it should.

Note that two of your tested implementations do this.  It is just that
one of them chooses to treat it as a non-fatal error in the .* case.
For the implementation that does not produce an error for the .name case,
you might want to see if it is creating an implicit net, which would be
even more wrong.

Steven Sharp
sharp@cadence.com