Cardinality reports one but should report two

[Peter Aronsson]

The following model reports cardinality(u) == 1 but should report cardinality(u) == 2:

model CardinalityBug
  
  Modelica.Blocks.Interfaces.BooleanInput u;
  Modelica.Blocks.Logical.And and1;
  parameter String card = "cardinality is "+String(cardinality(u));
equation
  connect(u,and1.u1);
  connect(u,and1.u2);
end CardinalityBug;

This language construct is used in Controlsystems in ThermoSysPro.

[Martin Sjölund]

[This is a deprecated operator. It should no longer be used, since it will be removed in one of the next Modelica releases.]

I'll have a (very) quick look.

[Martin Sjölund]

It's easy to fix this for this test in particular, but from what I can tell it breaks all the other code. (From my quick glance at the code, I suspect cardinality in OpenModelica only works correctly if there is at most one connect-equation using the cref in a particular component)

[Adrian Pop]

The problem is that we do union on the list when we should do list append in Ceval.cevalCardinality2. So per model we always get a cardinality of 1.
However, if you change that list union to list append then things start to
break as Martin pointed out (50 Models in the testsuite).
Why? Because we don't care about deleted components which are added later
on in the instantiation. So the actual problem is that we do not filter the
Env.CSetsType we add to the sets and to the env by the deleted components.

I will fix this myself. We should not Ceval.ceval the cardinality operator
at all during instantiation, only evaluate it at the end of instantiation
when we have all the information. Also we should either delete the connect crefs
matching deleted components from the Connect.SETS.connectionCrefs and the Env
or just use the info in the Connect.SETS as is the most updated one (I just have
to see how to get it from there in the way the cardinality wants it).

Cheers,
Adrian Pop/

[Per Östlund]

Fixed in r18987.