R devel - Nov 2003 - installing packages

If R_LIBS is not set then R CMD INSTALL without a lib argument installs  
in R_HOME/library
while install.packages() without lib installs in the first element of  
.libPaths().

If R_LIBS is set, then R CMD INSTALL installs in the
first directory in R_LIBS, and install.packages() still installs in the  
first element
of .libPaths() (i.e. ignores R_LIBS).

Is this difference intentional ? It would perhaps be more convenient if
.libPaths()[1] was used as a default in all cases (when no lib argument
is given).

==Jan de Leeuw; Professor and Chair, UCLA Department of Statistics;
Editor: Journal of Multivariate Analysis, Journal of Statistical  
Software
US mail: 8130 Math Sciences Bldg, Box 951554, Los Angeles, CA 90095-1554
phone (310)-825-9550;  fax (310)-206-5658;  email: deleeuw@stat.ucla.edu
homepage: http://gifi.stat.ucla.edu
   
------------------------------------------------------------------------ 
-------------------------
           No matter where you go, there you are. --- Buckaroo Banzai
                    http://gifi.stat.ucla.edu/sounds/nomatter.au

On Sun, 30 Nov 2003, Jan de Leeuw wrote:
> If R_LIBS is not set then R CMD INSTALL without a lib argument installs  
> in R_HOME/library
> while install.packages() without lib installs in the first element of  
> .libPaths().
which would normally be .Library=R_HOME/library, set in Rprofile.
> If R_LIBS is set, then R CMD INSTALL installs in the
> first directory in R_LIBS, and install.packages() still installs in the  
> first element
> of .libPaths() (i.e. ignores R_LIBS).
The first element of .libPaths() should be the first element of R_LIBS by 
default, set in Rprofile.
> Is this difference intentional ? It would perhaps be more convenient if
> .libPaths()[1] was used as a default in all cases (when no lib argument
> is given).
It is, provided the user has not reset .libPaths.

One caveat: R CMD INSTALL looks for R_LIBS in the environment, whereas by 
the time .libPaths() is set it can be set in other places, including via 
~/.Renviron.  Maybe R CMD INSTALL should look at more of those places.


-- 
Brian D. Ripley,                  ripley@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