Implicit iterator range and expandable arrays.

[qlambert-pro]

I was wondering if the following model is valid:

model AugmentArrayLoop
  expandable connector EC
  end EC;

  connector RealOutput = output Real;

  model M
    RealOutput ro = 1.0;
  end M;

  EC ec;
  M m;

  Real x[Boolean];
equation
  for i loop
    x[i] = 0;
    connect(m.ro, ec.ro[i]);
  end for;
end AugmentArrayLoop;

I couldn't find anything forbidding it, so I suppose it is.

[HansOlsson]

I agree that there is currently nothing forbidding it, but I believe it should be:

Specifically having an implicit iterator range based on an array in an expandable connector.
The reason is that the members of the expandable connector are also implicitly defined and combining two implicit definitions is not a good idea.

Note that implicit iterator ranges require that all indexed arrays have the same size, i.e. we cannot say that x determines the size and ec.ro just tags along.

The Boolean indexing of ec.ro seems odd, but I don't see any particular reason to forbid it.

[qlambert-pro]

Note that implicit iterator ranges require that all indexed arrays have the same size, i.e. we cannot say that x determines the size and ec.ro just tags along.

This is exactly what we do, and it works. we just disallow loops where the only usage of the iterator is as a subscript of a member of an expandable connector.

[HansOlsson]

Note that implicit iterator ranges require that all indexed arrays have the same size, i.e. we cannot say that x determines the size and ec.ro just tags along.

This is exactly what we do, and it works. we just disallow loops where the only usage of the iterator is as a subscript of a member of an expandable connector.

Ok, that would be possible but it feels odd to me, and that's not part of the spec:
https://specification.modelica.org/master/statements-and-algorithm-sections.html#implicit-iteration-ranges

If it is used to subscript several expressions, their ranges must be identical. The IDENT may also, inside a reduction-expression, array constructor expression, for-statement, or for-equation, occur freely outside of subscript positions, but only as a reference to the variable IDENT, and not for deducing ranges.

[HansOlsson]

Rewriting:

   for i in 1:size(y,1) loop
     y[i] = 0;
     connect(m.ro, ec.ro[i]);
  end for;

or:

   for i /* in 1:size(y,1) */ loop
     y[i] = 0;
   end for;
   for i in 1:size(y,1) loop
     connect(m.ro, ec.ro[i]);
  end for;

I thus see two possibilities:

• State that IDENT for implicit iteration ranges may not be used in subscript for arrays in expandable connector components. Martin, Filippo
• State that IDENT for implicit iteration ranges may be used in subscripts for arrays in expandable connector components, but only as a reference to the variable IDENT and not for deducing ranges. Quentin, Henrik, Adrian, Markus, Axel, Stefan

Go for 2nd option.

Gerd, Christoff, Dag, Hans abstain
Separately: And forbid/restrict size of array in expandable connector?