Inconsistency in treating array dimensions in modifiers

[modelica-trac-importer]

Reported by petfr on 18 Aug 2013 10:23 UTC
A design goal and property in Modelica is that array
dimensions should be possible either on the type or
on the component, i.e., both of the following are allowed:

Real[20] vec;
Real vec[20];

This language rule is broken for modifiers in the current version
of Modelica, MS 3.2.2, where the form vec[20] is allowed
but not Real[20]. This inconsistenly just recently caused
me a lot of confusion. The language grammar should be updated
to consistenly allow both forms.

Current grammar:
element_replaceable:
replaceable ( short_class_definition | component_clause1) [constraining_clause]

component_clause1:
type_prefix type_specifier component_declaration1

type_prefix: flow | stream | ...
type_specifier: name
component_declaration1: declaration comment
declaration: IDENT [ array_subscripts ] [ modification ]

Example which is accepted by current grammar and by Dymola:

package testredeclare

class C3
Real x1;
Real[:] vec;
end C3;

class C4 = C3(redeclare parameter Real x1, redeclare Real vec[20]);

end testredeclare;

Example which is not accepted by current grammar:

package testredeclare

class C3
Real x1;
Real[:] vec;
end C3;

class C4 = C3(redeclare parameter Real x1, redeclare Real[20] vec);

end testredeclare;

---

Migrated-From: https://trac.modelica.org/Modelica/ticket/1251

[modelica-trac-importer]

Comment by hansolsson on 24 Sep 2013 08:27 UTC
Language group:

Consider:

type Vec=Real[:];
type Vec2=Real[2];
class C3
  Real x1;
  Vec  x2[:];
  Vec [:] x3[:];
end C3;

Currently C3(redeclare Vec2 x2); means keep : from original declaration, i.e. Vec2 x2[:] and C3(redeclare Real x2[3,2]); is not allowed, since the "type" should have same dimensions as original declaration (end of section 7.3.2).

And C3(redeclare Vec x3[2]); is not allowed; i.e. the type-dimensions are not allowed to be "inherited" independently (end of 7.3.2). Either give all array dimensions or none.

So view Real[4] x[2] as syntactic sugar for Real x[2,4];; and either give all array dimensions or none; and thus logical to allow that in redeclare as well.

Conclusion: State that just syntactic sugar and allow in redeclare as well.
Agreement.
Poll: Include in 3.3r1?
Agree: 0
Against: 1
Abstain: 5
Conclusion: needs more work.

[modelica-trac-importer]

Comment by hansolsson on 1 Mar 2017 14:11 UTC
In 7.3 state that:
Replaceable component array declarations with array sizes on the left of the component are seen as syntactic sugar for having all arrays sizes on the right of the component; and thus can be redeclared in a consistent way.

Example:

  model M
    replaceable Real [4] x[2];
  end M;
  M m(redeclare Modelica.SIunits.Length x[2,4]);

Agreement by acclamation.

[modelica-trac-importer]

Comment by hansolsson on 14 Mar 2017 10:34 UTC
Resolved in r9681