[Rd] Recursive dependencies(Rcmdr)

[John Fox]

Dear Brian and Peter,

I guess that this problem didn't surface before because most use of the
Rcmdr package is on Windows using the package binaries.

I think that I can simply remove mvtnorm (and probably some others as well)
from the Rcmdr dependencies. I'll check more carefully when I have some
time, but I believe that mtvnorm is there just for multcomp, and its
inclusion probably dates to a period before dependencies were resolved by
install.packages().

Thanks for bringing this to my attention.

John

--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox 
-------------------------------- 

> -----Original Message-----
> From: r-devel-bounces at r-project.org 
> [mailto:r-devel-bounces at r-project.org] On Behalf Of Prof Brian Ripley
> Sent: Friday, November 11, 2005 10:45 AM
> To: Peter Dalgaard
> Cc: r-devel at stat.math.ethz.ch
> Subject: Re: [Rd] Recursive dependencies(Rcmdr)
> 
> On Fri, 11 Nov 2005, Peter Dalgaard wrote:
> 
> > Something might have slipped by me, but I got into the following 
> > situation installing Rcmdr:
> 
> install.packages() does indeed have code to check for 
> dependencies of dependencies.  I got the message
> 
> also installing the dependencies 'acepack', 'scatterplot3d', 
> 'quadprog', 'fBasics', 'Hmisc', 'mlbench', 'randomForest', 
> 'SparseM', 'xtable', 'oz', 'leaps', 'dynlm', 'e1071', 
> 'tseries', 'chron', 'fCalendar', 'its', 'DAAG', 'abind', 
> 'car', 'effects', 'lmtest', 'multcomp', 'mvtnorm', 'relimp', 
> 'sandwich', 'strucchange', 'zoo'
> 
> The problem is the ordering.  Rcmdr does have mvtnorm in the 
> 'Suggests' 
> list, and has it after 'multcomp', and install.packages does 
> not check that the Suggests list does not have inter-dependencies.
> 
> So we do `do something smarter', but not smart enough.
> 
> >
> > < install.packages("Rcmdr",depend=TRUE) >
> >
> > Oodles of Output, until:
> >
> > * Installing *source* package 'multcomp' ...
> > ** R
> > ** data
> > ** inst
> > ** preparing package for lazy loading
> > Error in dyn.load(x, as.logical(local), as.logical(now)) :
> >        unable to load shared library 
> '/home/bs/pd/Rlibrary/mvtnorm/libs/mvtnorm.so':
> >  /home/bs/pd/Rlibrary/mvtnorm/libs/mvtnorm.so: cannot open shared 
> > object file: No such file or directory Execution halted
> >
> > .... Several oodles more for remaining packages ....
> >
> > ** building package indices ...
> > * DONE (strucchange)
> >
> > The downloaded packages are in
> >        /tmp/RtmpF23254/downloaded_packages
> > Warning messages:
> > 1: installation of package 'multcomp' had non-zero exit status in: 
> > install.packages("Rcmdr", depend = TRUE)
> > 2: cannot create HTML package index in: 
> > tools:::unix.packages.html(.Library)
> >
> >
> > OK, so we start Rcmdr
> >
> >> library(Rcmdr)
> > Loading required package: tcltk
> > Loading required package: car
> >
> > it then detects that multcomp is missing and asks for permission to 
> > install it, giving the same result.
> >
> > The thing appears to be that multcomp depends on mvtnorm, but the 
> > extra dependency goes undetected. Manual installation of the two 
> > works.
> >
> > A workaround is that John adds mvtnorm to the Suggested list for 
> > Rcmdr, but I wonder if we couldn't do something smarter.
> >
> >> version
> >         _
> > platform x86_64-unknown-linux-gnu
> > arch     x86_64
> > os       linux-gnu
> > system   x86_64, linux-gnu
> > status
> > major    2
> > minor    2.0
> > year     2005
> > month    10
> > day      06
> > svn rev  35749
> > language R
> >> .libPaths()
> > [1] "/home/bs/pd/Rlibrary" "/usr/lib64/R/library"
> >
> > -- 
> >   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
> >  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
> > (*) \(*) -- University of Copenhagen   Denmark          Ph: 
>  (+45) 35327918
> > ~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  
> FAX: (+45) 35327907
> >
> > ______________________________________________
> > R-devel at r-project.org mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-devel
> >
> >
> 
> -- 
> 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