Reading the r-devel changes RSS feed has its benefits as I noticed the other day (Feb 4 to precise) that

CHANGES IN R-devel NEW FEATURES
‘qnorm(, log.p=TRUE)’ is now fully accurate (instead of to “only” minimally five digits).

And indeed, we now get

  ----- FAILED[data]: test_stats.R<246--246>   call| expect_equal(runit_qnorm_log(-1e+05)$lower, -447.1974945)   diff| Expected '-447.1974945', got '-447.197893678525'  Error: 1 out of 1590 tests failed

Fix is easy and I'll push it once the CITATION issue in#1250 is dealt with.