I introduced the doRNG package to ensure reproducibility on Windows.

Great @wiesenfa! The test with doRNG passed on Windows and Ubuntu!

I implemented Manuel's suggestions in branch hotfix/T29361-EnsureReproducibilityWithParallelBootstrapping and added corresponding unit tests to test-bootstrap.R.

I like the results when the scales library is used! However, when we find a way to bring back the confidence intervals, also @wiesenfa's latest solution can be used.

First, I tried the fix with R 3.6 and can confirm that it does not break the functionality there.

Could someone with R v4 double-check locally as well?

I checked on my local Windows system (R 3.6.3) and the plots seem to be correct.

I have tested this with the provided data. The scaling of the y-axis seems to be correct now. But only the first rank is labeled on the y-axis. Can the other ranks be labeled as well?

Thank you for investigating this! In challengeR it is covered in the way that a message is emitted saying "na.treat obligatory if report is intended to be compiled". In order to solve the mentioned issue 2, a strategy for the preferred way to handle it in VISSART should be defined. Should the user be guided to specify the NaN handling strategy? Should the user be able to generate a report but without the plots that require numeric values?

One solution to approach this issue is to check if NaN values are present in the data. If so, the user should be able to select the NaN handling strategy if necessary (e.g., not for case-based ranking). So what about providing the NaN handling options after the chosen ranking method is known?

Successfully checked the proposed fix on a Windows system for downward compatibility (R 3.6.3, ggplot2 3.3.0).