Francesco Casella

> When investigating this model it became immediately clear why the equation is generated as it is. We only remove connectors from the set when dealing with `inStream`, but not...

LGTM. BTW, stream connectors are extensively used in the Buildings library, for which we are carrying out verification against Dymola-generated results, so if there's something wrong, chances are that the...

In fact, I had two further comments, but had no time to write them down until now. The first, you already anticipated: the condition to remove stream connectors from the...

As I mentioned in #9302, print statements in equation sections are a bit fishy. What about putting them within an algorithm?

@kabdelhak, @AnHeuermann, there is something fishy going on here with causality analysis, concerning stand-alone function calls. Can you please check @max-privato's test case and see what's wrong?

Thanks @AnHeuermann. In my experience this kind of error often corresponds to a singular Jacobian. Whether this is due to bad start values or to structural issues in the backend,...

Doesn't LOG_NLS_JAC do the trick? I used it in the past, but not with Kinsol, only with the internal hybrid solver

I'm not going to read that, Matlab's going to 😅

@ceraolo, one side comment before getting to the real issue: the attached package has a uses(Modelica) annotation in _the model_ and not _in the package itself_, which doesn't really work....

@ceraolo thanks for reporting, this is really a weird bug. I checked with the declarative debugger, and the parameter seems to be properly propagated: ``` 170 parameter (assign) drive.limTau.wMax :=...