Open modelica-trac-importer opened 6 years ago
Comment by hansolsson on 24 Sep 2013 10:18 UTC
Language group:
Would require a grammar change for constrainedby Real[:]
.
It becomes complex if the dimension is not required on constrainedby, and incompatible if required on constrainedby.
It would also allow you to define a dimensions as changeable with a default by making the component replaceable which seems like a new awkward way (currently it is possible by using a parameter for dimension which is also awkward).
The conclusion is that a more complete redesign is needed and not clear. Postpone.
Having dimensions in the constrainedby seems necessary. For example I've seen something weird like this:
replaceable IBPSA.Fluid.MixingVolumes.BaseClasses.MixingVolumeHeatPort vol[nEle]
constrainedby IBPSA.Fluid.MixingVolumes.BaseClasses.MixingVolumeHeatPort(
redeclare each package Medium = Medium,
each energyDynamics=Modelica.Fluid.Types.Dynamics.FixedInitial,
final initialize_p={(i == 1 and not Medium.singleState) for i in 1:nEle},
each m_flow_nominal=1,
each V=1,
each nPorts=2) "Mixing volume";
which is quite wrong as you have array modifiers on a non array class in constrainedby, but Dymola accepts it fine. It would be correct if written as:
replaceable IBPSA.Fluid.MixingVolumes.BaseClasses.MixingVolumeHeatPort vol[nEle]
constrainedby IBPSA.Fluid.MixingVolumes.BaseClasses.MixingVolumeHeatPort[nEle](
redeclare each package Medium = Medium,
each energyDynamics=Modelica.Fluid.Types.Dynamics.FixedInitial,
final initialize_p={(i == 1 and not Medium.singleState) for i in 1:nEle},
each m_flow_nominal=1,
each V=1,
each nPorts=2) "Mixing volume";
Currently the only way to deal with this properly is using a short class definition:
model T = IBPSA.Fluid.MixingVolumes.BaseClasses.MixingVolumeHeatPort[nEle];
replaceable T vol
constrainedby T(
redeclare each package Medium = Medium,
each energyDynamics=Modelica.Fluid.Types.Dynamics.FixedInitial,
final initialize_p={(i == 1 and not Medium.singleState) for i in 1:nEle},
each m_flow_nominal=1,
each V=1,
each nPorts=2) "Mixing volume";
Reported by petfr on 16 Aug 2013 06:25 UTC In Section 7.3.2 of MS 3.2 rev 2 we have the following:
... The number of dimension in the constraining type should correspond to the number of dimensions in the type-part and in the constraining type.. Similarly the type used in a redeclaration must have the same number of dimensions as the type of redeclared element. [Examples: replaceable T1 x[n] constrainedby T2; replaceable type T=T1[n] constrainedby T2; replaceable T1[n] x constrainedby T2; In these examples the number of dimensions must be the same in T1 and T2, as well as in a redeclaration. Normally T1 and T2 are scalar types, but both could also be defined as array types – with the same number of dimensions. Thus if T2 is a scalar type (e.g. type T2= Real) then T1 must also be a scalar type; and if T2 is defined as vector type (e.g. type T2=Real[3]) then T1 must also be vector type.] ...
The simple intuitive rule regarding constraining type is that it gives a type constraint (supertype) regarding a replaceable type or type of a replaceable component, e.g.:
replaceable type Mytype=T constrainedby T2; replaceable Mytype x constrainedby T2;
In both cases Mytype must be a subtype of T2.
According to the same reasoning, the following should be correct, since Real[2] is a subtype of Real[:]
replaceable type T=Real[2] constrainedby Real[:]; replaceable Real[2] x constrainedby Real[:];
But according the the above text in MS 3.2 rev 2 with T1=Real and T2 = Real, it should be as follows, with a special rule for the dimensions of a type:
replaceable type T=Real[2] constrainedby Real; replaceable Real[2] x constrainedby Real;
This is inconsistent and contradictory language design and is a prime example of a rule that makes the language hard to understand and learn.
Migrated-From: https://trac.modelica.org/Modelica/ticket/1247