Thanks, it is all starting to make sense now!

This example shows why the output_expression_list is actually useful:

model Test



function func

  input Real a;

  input Real b;

  output Real c;

  output Real d;

algorithm

  c := a*a;

  d := b+b;

end func;



  parameter Real e = 3;

  parameter Real f = 5;

  Real g;

equation

  (,g) = func(e,f);

end Test;

I'll comment this also to the bug report.

Modelica specification has a separate definition for that case:

statement :

( component_reference ( ":=" expression | function_call_args )

| "(" output_expression_list ")" ":=" component_reference function_call_args 

| ...

Now when thinking of this, replacing output_expression_list with expression_list would allow (1,2,3)^(4,5,6,7,) which would not make much sense either.. Maybe that is just a forgotten legacy line for the function call case? I'll add that ticket and wait for smart people to resolve it

Is there anyone who could explain why Modelica syntax allows "(" output_expression_list ")" for primary?

Here is a clip from Modelica 3.3 spec (https://www.modelica.org/documents/ModelicaSpec33.pdf)

primary :

UNSIGNED_NUMBER

 | STRING

 | false

 | true

 | ( name | der | initial) function_call_args

 | component_reference

 | "(" output_expression_list ")"

 | "[" expression_list { ";" expression_list } "]"

 | "{" function_arguments "}"

 | end

The problem is that output_expression_list allows expressions like (,,,,,,)^(,,,,,). They go unnoticed to compiler which obviously cannot process them. Should it be expression_list instead? Just wondering

So I was a little ahead of myself. Too bad, I was so excited about this

What kind of plans there are to support FMI 2.0? I suppose FMI 2.0 will be released pretty soon, since it was planned for March 2012.

Hello all!

I have modified FMU SDK http://www.qtronic.de/en/fmusdk.html and use it to run models exported as .fmu using OpenModelica. During my testing, I ran into a problem, which is more of a philosophical problem than a software problem. Let's say I have the following model:

model FMITest

    Real stock(start=0.0, fixed=true);

    parameter Real par = 4;    

    parameter Real valve = par;

equation

    der(stock) = valve;

end FMITest;

The export produces model description containing:

  <ScalarVariable

    name="par"

    valueReference="2"

    variability="parameter"

    causality="internal"

    alias="noAlias">

    <Real start="4.0" fixed="true"  />

  </ScalarVariable>

  <ScalarVariable

    name="valve"

    valueReference="3"

    variability="parameter"

    causality="internal"

    alias="noAlias">

    <Real start="4.0" fixed="true"  />

All good in there. But let's say that during simulation, I change the value of "par" to 2. In my case, I would like that the value of "valve" is also changed. Currently this does not happen and there is no alias reference between"par" and "valve", which I could follow in my software. Is there a specification that says this is the way the software should work or could it be changed? I know that leaving the "parameter" keyword out from "valve" and specifying "valve = par" in equation block would probably fix this, but I would like to leave my model as it is for other reasons.

-Teemu

Thanks for the fast and good reply! I got rid of +showErrorMessages, since I clearly don't need it.

-Teemu

Hi!

After installing OpenModelica 1.7.0 RC1 I get following errors when using simulate or buildModel in .mos -files in windows command prompt:

{"Error: Class OpenModelica.Scripting.simulate not found in scope <global scope> (looking for a function or record).", "TRANSLATION", "Error", "3"}

Example .mos -file:
loadFile("testi2.mo");
simulate(Model1, startTime=0,stopTime=1, numberOfIntervals=10, method="euler", outputFormat="plt");
getErrorString();

The model is simulated even though the error is shown, but I would like to get rid of the error message. What could be the reason for the errors?

Thanks!

-Teemu