[R] Oddities upgrading from 2.4.1 to 2.5.0

[Prof Brian Ripley]

You are confusing binary packages available for a version of R for 
Windows, and source packages.  But 'stats4' is not on CRAN.

On Sun, 13 May 2007, Chris Evans wrote:

> I'm working on Windoze XP and have for some time installed R in D:\R and
> upgraded my packages after installing the new version of R with a bit
> of code I think I got from the list:
> ip <- installed.packages(lib.loc = "d:\\R\\R-2.4.1\\library")
> ip <- ip[ip[,1]!="base" ,1]
> install.packages(ip)
>
> This time around that's produced some warning grumbles:
> Warning messages:
> 1: packages 'dynamicGraph', 'ggplot', 'grid', 'Qtoolbox', 'ref',
> 'splines', 'stats4', 'tapiR', 'tcltk', 'tools' are not available

"grid", "splines" "stats4", "tcltk" and "tools" come with R.  Your 
selection test is wrong (you want to test for NA Priority).
Something like

need <- as.vector(ip[is.na(ip[, "Priority"]), 1])

> 2: dependencies 'Design', 'graph', 'RBGL' are not available

"graph" and "RBGL" are on Bioconductor, and using 'Select repositories' 
from the menu to add that repoository would have found them.

For "Design", "dynamicGraph", "ggplot", "ref" see

http://cran.r-project.org/bin/windows/contrib/checkSummaryWin.html

CRAN does not distribute binary versions of packages that fail their 
tests.

AFAIK tapiR has been withdrawn from CRAN.


> Sure enough, those in warning 1 were all in the 2.4.1 library directory
> tree and those in warning 2 weren't.  Now I'm trying to work out how
> these failures have happened.
>
> I think some of this could be because of a problem with the mirrors
> (though I've tried both UK(London) and Switzerland(Zurich)).  I have
> seen the posting saying that there are problems.

Not at all.

> I also think some could be because I may have installed some packages 
> under 2.4.1 which didn't come from CRAN but I didn't keep careful 
> records stupidly.
>
> Anyway, looking at http://cran.r-project.org/ packages I think that:
> 'dynamicGraph', 'ggplot', 'ref', 'stats4' and 'Design' should all be
> there for me to install with install.packages() but they're definitely
> not at UK(London) or Switzerland(Zurich).

You are not looking at the right list on that website.

> Searching a bit I can see that Qtoolbox was a non-CRAN package I
> installed directly.
>
> However, I still seem to have created some warnings here that searching
> the R site search isn't answering for me.
>
> Two questions/suggestions:
> 1) is there a list somewhere of packages/libraries that were once in R
> but aren't now?
> 2) is there a way of checking for missing dependencies that will
> identify both what's missing (clearly something like this is producing
> warning 2 following my call of "install.packages(ip)") but might also
> produce a list of the origins of the dependencies?

It _is_ in the value returned by available.packages(), and in that 
returned by installed.packages() (but dependencies can change with later 
versions than those you have installed).

> Sorry if these are trivial issues.

-- 
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