Problems with daeMode in OMEdit

[Francesco Casella]

Consider the attached simple model. If I run the .mos script from the command line, everything works, and I get a proper .mat file with the expected results.

If I load the .mo file in OMEdit and run it from there, the model is compiled, the simulation output window shows me a 100% green simulation bar, it tells me "Simulation of TestDAEMode is finished", but I don't see the final statistics in the window, the .mat file is not generated, and no results are visible in the Plotting tab.

Can you please fix this? I'm not sure whether the problem is with DAEmode or with OMEdit, though I tend to think that OMEdit has some issues, as the command line version works fine.

[Willi Braun]

Ah, it seems to be an issue in the runtime. Now the only solver that is possible to use in DAEmode is the IDA solver, I stopped the support of dassl in DAEmode.
Actually the runtime should select directly IDA as solver when the model is compiled in DAEmode, but this seems to conflicting with the override arguments. I'll check it.

[Francesco Casella]

In fact, beyond the probles reported in comment:1, there is a more fundamental problem with the vendor annotation. I opened #4812 about that.

[Adeel Asghar]

I got the model to work fine with the following change,

model TestDAEMode
  Real x(start=0,fixed=true),y;
 
equation
  der(x) = y;
  y= -x+1;
  annotation(experiment(StopTime = 5, Tolerance = 1e-6),
     __OpenModelica_simulationFlags(daeMode = ()),
     __OpenModelica_commandLineOptions = "--daeMode=all");
end TestDAEMode;

Note that I added () to daeMode flag.

I have written a small documentation about __OpenModelica_simulationFlags (​https://github.com/OpenModelica/OpenModelica-doc/pull/72) which should be up soon.

[Adeel Asghar]

The documentation is available now ​https://openmodelica.org/doc/OpenModelicaUsersGuide/latest/omedit.html#openmodelica-simulationflags-annotation