Ralf Stubner

@jlmelville Please use the state from #79 for testing. Current main is missing an important `inline`. Concering `dqsample.h`: It is best to wait until I have done a CRAN release...

@jlmelville You could upload anytime, but it probably makes sense to wait until I have finalized #79. Feel free to add me as contributor if you see this as important...

@s3alfisc: I was able to reproduce the issue with the CRAN version 0.13 of `fwildclusterboot`, while both the current development version as well as version 0.14.3 from r-universe does not...

@joemsong: I was able to reproduce the issue with the CRAN versions of both `GridOnClusters` and `FunChisq`. In both cases, it is enough to insert `dqrng::dgRNGkind("xoroshiro128+")` in the relevant test...

@ahudde: I was able to reproduce the issue with the CRAN version of `greeks`. It is enough to insert dqrng::dgRNGkind("xoroshiro128+") in the relevant test file. However, in the current development...

Thanks @ahudde and @s3alfisc for looking into this! A CRAN upload of your packages in the coming weeks with an appropriate fix would be great.

@jlmelville Yes, please go ahead.

:tada: Latest [revdep check](https://github.com/daqana/dqrng/actions/runs/8750447939) no longer shows a failure with `rnndescent`. Interestingly, `fwildclusterboot` was fine as well, even though this is still the original version.

@s3alfisc: It is odd, that `fwildclusterboot` does not show any problems in the last revdep check. And looking at the [previous results](https://github.com/daqana/dqrng/actions/runs/8701720789/job/23865004851) it was actually the CRAN version of dqrng...

@jlmelville This is strange. I am currently seeing reverse dependency failures from `rnndescent` again. This time not from the installation but from the tests in `test_rptree.R`. Do you know what...