Taking up the thread that Ben Bolker started on Feb.5 ... In the mean time, with lots of contributions from Petr Savicky, we have fixed the bug (and even related bugs that were not reported).
The consequence is that a sizable fraction of all (numeric) R outputs does change (very) slightly, typically using a digit more or less than previously. Several of the recommended packages, e.g., needed some adjustments of their tests/*.Rout.save files. Here are the current NEWS entries related to that: SIGNIFICANT USER-VISIBLE CHANGES: [ ........ ] • Printing and formatting of floating point numbers is now using the correct number of digits, where it previously rarely differed by a few digits. (See “scientific” entry below.) This affects _many_ *.Rout.save checks in packages. NEW FEATURES: [ ........ ] • The internal (C) function scientific(), at the heart of R's format.info(x), format(x), print(x), etc, for numeric x, has been re-written in order to provide correct results, fixing PR#14491, notably in border cases, including when digits >= 16, thanks to substantial contributions (code and experiments) from Petr Savicky. This affects a noticable amount of numeric output from R. --------- Feedback is welcome, but flames > /dev/null Martin Maechler ______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel