Changes in r-devel last week lead to an email to Rcpp (see [this issue](https://github.com/RcppCore/Rcpp/issues/1308)) also mentions new UBSAN issues for `nanotime` :crying_cat_face: See [this log](https://www.stats.ox.ac.uk/pub/bdr/memtests/gcc-UBSAN/nanotime/tests/tinytest.Rout) which I haven't tried to replicate yet.