Paul,
It so happens that I know the R stack pretty well, and I happen to know how CRAN operates. Getting a new upstream package from CRAN almost always means that nothing is broken, or that (a rare case) a change affects a downstream package -- as all this is tested well at CRAN. I am pretty close to that ecosystem, and there is mutual trust between them and me. Hence an autopkgtest failing was likely an indication of _our_ downstream dependencies of this package failing. And so it was, and that I demonstrated. By now both these packages have been updated https://packages.debian.org/sid/r-cran-lmertest https://packages.debian.org/sid/r-cran-mlmrev and there is still no bug in my package r-cran-lme4. We "merely" have to wait for testing to receive the newer packages. And then r-cran-lme4 will migrate. Until there is nothing for me to day, just as there was nothing for me to five days ago. It really was an issue in r-cran-{lmertest,mlmrev}. Hope this helps. Dirk -- http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org