Francesco Casella

Shall I open a ticket on the Buildings repo or can you take care of that?

@hcasperfu according to the Modelica Language Specification, [Section 15.3](https://specification.modelica.org/maint/3.6/stream-connectors.html#actualstream): > Since the output of actualStream will be discontinuous, it should not be used by itself to model physical behaviour (e.g.,...

@phannebohm, @kabdelhak, can you please check?

I wonder if it really makes sense to expect that the simulation of switching systems gives exactly the same results no matter how the numerical solution is carried out. I...

To me the case is clear: the exact behaviour of this signal is very sensitive to numerical errors, but the physical outcome (e.g. output temperatures) isn't, since the other signals...

@HansOlsson I'm not sure about that. It is [well-known](https://mbe.modelica.university/behavior/discrete/decay/) that writing something like ```modelica v = if x>=0 then sqrt(x) else 0; ``` does not prevent a tool from computing...

I checked the structure of those failing models. They all have one feature in common: on the left side, there is a V_pulse component which is obviously meant to generate...

Comments from other tool developers are welcome 😃

I checked more in depth what happens in OpenModelica. If we define `ModelicaServices.Machine.inf` with any value up to half than the largest representable IEEE 754 double precision number (0.89885e308) this...

In fact, it turns out the issue with OpenModelica was triggered by #4042, which changed https://github.com/modelica/ModelicaStandardLibrary/blob/d2dcb39e055ceb88fef5328a26255cdd094389bd/ModelicaServices/package.mo#L185-L186 into https://github.com/modelica/ModelicaStandardLibrary/blob/8e7876b5414475abab003c466ca49050812aeb01/ModelicaServices/package.mo#L184-L185 so in some sense it is indeed a regression caused by changes...