Re: base R packages conflicting with CRAN packages

[Jason Bacon <bacon4000%gmail.com@localhost>]

On 02/27/17 11:37, Kamil Rytarowski wrote:
> On 25.02.2017 05:21, Brook Milligan wrote:
> > I have identified a case in which the math/R package installs a recommended package (in this case survival v2.39.5) that is not a new enough version to satisfy the needs of another package, which needs survival v2.40.1.  Because survival is installed with the base R package it is impossible to replace it with a newer R-survival package because they would inevitably conflict.  This is not a case of R being old, as pkgsrc installs the most recent release.  Instead, it is a case of the version of survival recommended by base R being older than the most recent survival package available from CRAN.
> >
> > How should this be handled?  Must we turn all the packages otherwise installed by base R into individual packages so that they can be upgraded as needed independently of base R?  If so, how do we prevent pkgsrc from installing the recommended set of packages along with R?What other options are there?
> >
> > Thanks for your input.
> >
> > Cheers,
> > Brook
> It is off-topic but I gave up on R packages due to lack of PLIST files.
> This made my work on R uncontrollable and unmaintainable. That said
> probably it's just better to rely on packages from CRAN directly without
> pkgsrc...

I've been leaning in this direction as well.  It seems impossible to 
keep the R-cran packages in any package manager up-to-date and 
complete.  Our users inevitably have to build some of their own, which 
is easy within an R script.

This can become a bottleneck when using the official cran mirrors from a 
parallel computing grid (with no shared home directory), but it's easy 
to create a local mirror:

https://cran.r-project.org/mirror-howto.html

--
Earth is a beta site.