[R] Cleaning up R-2.0.0 installation on GNU/Linux ?[SOLVED]

[Prof Brian Ripley]

On Wed, 13 Oct 2004, Thomas Schönhoff wrote:

> Hello,
> 
> 
> Prof Brian Ripley said the following on 13.10.2004 10:34:
> > On Wed, 13 Oct 2004, Thomas Schönhoff wrote:
> 
> > remove.packages("nlme", "/usr/local/lib/R/site-library") # a duplicate
> > foo <- installed.packages(lib="/usr/local/lib/R/site-library")[, 1]
> > install.packages(foo, lib="/usr/local/lib/R/site-library")
> 
> Besides one minor error regarding to distr package:
> 
> --------------------------------------------------------------
> Warning message:
> Installation of package distr had non-zero exit status in: 
> install.packages(foo, lib = "/usr/local/lib/R/site-library")
> ---------------------------------------------------------------

You seem to be missing its dependency setRNG, I should have suggested

install.packages(foo, lib = "/usr/local/lib/R/site-library", 
                 dependencies = TRUE)

but thought about it and thought that you must have all the dependencies 
if things are working in 1.9.1 (so presumably they were not).

> your recommendations worked out fine, thanks!
> 
> 
> 
> > will reinstall all the packages in that library.  I could work out a way
> > to only get those that were installed in an earlier version, but it would 
> > be cumbersome.
> 
> No need for doing this at the moment!
> 
> 
> If understand correctly the previous situation is only arising when 
> there are major version upgrades, i.e. from 1.9.1 too 2.0.0.

It has only happened once in R's history.

> But when using "update packages() within a major release everything will 
> be fine, isn't it?

Yes, but even when upgrading a minor version is may be a good idea to 
re-install as, e.g. the processing of help files may get improved or more 
metadata may be installed.

Maybe we should add an option to update.packages to reinstall old 
installations too.

> I would be interested to hear what upgrade methods GNU/Linux (Debian) 
> users prefer to avoid a situation like this!
> 
> Many thanks for help!
> 
> Thomas
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
> 
> 

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595