On 20/04/2014, 6:43 PM, Jeroen Ooms wrote:
On Sun, Apr 20, 2014 at 11:19 AM, Martin Maechler
<maech...@stat.math.ethz.ch> wrote:
you are welcome, and thanked again together with everyone who
spends the little extra time for checking / using "the next
version of R" -- in general, i.e., on a regular basis.
With the risk of starting another flamewar I would like to point out
that a natural way to encourage use of r-patched/r-dev would be
extending the R release cycle to CRAN packages as was proposed on this
mailing list a couple of weeks ago.
If install.packages would default to downloading frozen/stable CRAN
packages in r-release, and download the very latest CRAN packages in
r-patched/r-dev, then most developers would probably switch to r-dev.
I doubt if that would be true. I'd guess most everybody would use
r-patched. This would have the effect that nobody would have
reproducible research (since r-patched builds aren't archived), and few
would test r-devel.
Duncan Murdoch
Thereby they would automatically be using the latest ("devel")
versions of base packages in the same way as they currently test their
script/project/package to work with the latest ("current") versions of
CRAN packages.
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
