Francesco Casella

That's precisely what I mean, this formulation is fishy 😄

It's a bit like writing ```C a = ++b + (c = d / b) – 1; ``` in C. For sure it is legal C code, but why would...

Yeah, what I mean is that the initialization with the start attribute is a bit tricky and not very easily readable.

> By allowing multiple versions of a library to be loaded, you could simply keep on using libraries without updating them. But it would be quite confusing to a user...

Based on the discussion so far, could we agree on the following scenario: - We make it more explicit that when flattening a model, only one version of each library...

I can't see the results of the testsuite. Maybe you can run it on this PR, so we can check?

> > > @casella @ptaeuber I couldn't remember the discussion on #1668. How should we deal with it? As I understand, the discussion there refers to variables that have some...

@lochel > There are two models in the testsuite that give bad results: > > * simulation/libraries/3rdParty/ThermoSysPro/ThermoSysPro.Examples.SimpleExamples.TestStodolaTurbine.mos > > * simulation/Modelica/inlineFunction/inlineArray1.mos I checked the [Hudson report](https://test.openmodelica.org/hudson//job/OpenModelica_TEST_PULL_REQUEST/6736/), there are 32 failure...

> Exactly. Most of the test do no longer show the warning about iteration variables without start values. Well, this seems in fact a very positive outcome, isn't it? 😄...

> * inlineArray1.mos is not compiling anymore I had a closer look at the BLT this test case. Vector `z` is only function of time, so it can be computed...