Sync comparisonSignals.txt with MSL d2592ead8b

[beutlich]

Sync the comparisonSignals.txt with modelica/ModelicaStandardLibrary@d2592ea.

Note that MSL is seen as SSoT. I wonder why they differ at all.

[maltelenz]

From what I can tell, there are 4 categories of differences:

1. Whitespace changes
2. Reordering of the same variables
3. The copy from ModelicaStandardLibrary has variables removed, but those variables still exist in this repo (both in comparisonSignals.txt and in the actual csv files.
4. New files that don't have any references yet

I think 1 and 2 are unfortunate, and it could be helpful if LTX improved their tooling to not modify the comparisonSignals.txt they create, just taking the ones from ModelicaStandardLibrary.

3 is an effect of the decision to not update references, when variables are removed from the set to be tested, but no new variables are added. I'm not sure if we should do anything for this repository in this case?

4 is just that modelica/ModelicaStandardLibrary#4349 hasn't been finished yet, and introducing the comparisonSignals.txt for model we don't have reference results for yet seems like a bad idea.

In the end, I begin to question why this repo contains comparisonSignals.txt to begin with?

[beutlich]

Thanks for the thorough wrap-up.

In the end, I begin to question why this repo contains comparisonSignals.txt to begin with?

When LTX or myself run the regression tests initially, it created all output files in the same folder where the comparisonSignals.txt are stored. So we copied the entire folder structure and created this repository including the comparisonSignals.txt (which got duplicated).

I'd expected the same approach for the subsequent regression test runs (like I setup for modelica/ModelicaStandardLibrary#4359 (comment)), hence I wonder why the comparisonSignals.txt ever can get out-of-sync.