Thomas Beutlich

> but I don't see that this PR currently solves the issue. Not sure what you mean exactly. Can read it as the patch is not good enough or not...

As you can imagine I am not happy with this rejection. The reason is that now the second derivative no longer corresponds to the first derivative: Jump to zero in...

I am struggeling with the (deduced) unit of `s`. If the start value `fluidLevel_max` is of type `SI.Height` ist should be a length based unit, right?

What about `SI.Length[nPorts] s(each start = fluidLevel_max)` to make it explicit and to not deduce the min attribute of SI.Height of the start value?

In a similar pattern: ModelicaTest.Media.TestAllProperties.IncompleteMedia.ReferenceMoistAir prints 1 invalid signal out of 27 on the overview page, but it actually is 3 out of 27 when you open the comparison report....

I am afraid I cannot confirm your observations, i.e., the `Failed values` are listed as expected. Find my log files attached: [ReferenceMoistAir.zip](https://github.com/modelica/ModelicaStandardLibrary/files/14923082/ReferenceMoistAir.zip). You can see that I obtain different values...

@MatthiasBSchaefer csv-compare [v2.0.4](https://github.com/modelica-tools/csv-compare/releases/tag/v2.0.4) fixed the need for counting failed values by printing the total number.

> Please use an other regression testing tool if you need this ! For the record, these are the existing OSS testing frameworks I am aware of: * https://github.com/modelica-tools/msl-release *...

> @MatthiasBSchaefer csv-compare [v2.0.4](https://github.com/modelica-tools/csv-compare/releases/tag/v2.0.4) fixed the need for counting failed values by printing the total number. @MatthiasBSchaefer I wonder if LTX ReSim uses csv-compare v2.0.4. It' closed source and I...

Edit, there also is #4372. I thought that maint/4.0.x no longer gets updates.