[R] Different missing links on Windows in 'check' vs. 'install'

[Robert Gentleman]

On Tue, Mar 09, 2004 at 12:38:33PM +0000, Prof Brian Ripley wrote:
> On Tue, 9 Mar 2004, Duncan Murdoch wrote:
> 
> > On Tue, 9 Mar 2004 08:03:44 +0000 (GMT), you wrote:
> > 
> > >No, both will find links in the same library as installing into (plus
> > >those which are fixed up on installation, e.g. to the base package).
> > >
> > >Several of us have looked for years for a fix, and this is the best scheme 
> > >we have come up with.  You can't put in absolute paths in the HTML as e.g. 
> > >a private library may be used with more than one version of R (or R may be 
> > >updated later).  Short of adding symbolic links to Windows (and getting 
> > >browsers to follow them), how do you propose `we should fix' it?
> > 
> > Here's a proposal:
> > 
> > All of check and build and install should default to the same library
> > location.  Check and build aren't meant to be permanent installs, so
> > if the package already exists there, it'll have to be temporarily
> > moved out of the way.
> 
> You may not own the main library and so not have permission to 
> install/check/build there.  If you do, you can use check on an installed 
> copy of the package.  As for build, this is one of the reasons why
> Rcmd INSTALL --build was needed, as that installs in the standard place 
> and then wraps up the installed copy.

 Yes, but for automated checking of many packages (60+ in Bioconductor
 and over 200 in CRAN) it would be nice to be able to separate out the
 real issues from the supposed ones.

 Perhaps we need a mechanism to generate some sort of database of
 these (which is updated/modified by R INSTALL, for example). Then R
 CMD check could generate an index file of resolved/supplied 
 links for the library path that R is using, generally
 $R_HOME/library, and then we simply compare against that. It seems
 that we are currently making the wrong comparison. What might
 make more sense (it seems to me) is to check and see if the  links
 are resolved only against base R + any packages in the
 depends/suggests field of the package being  checked. Since these are
 the ones that we will try to ensure are  available at run
 time. References outside of that set could be warnings. 

 Here we are not trying to physically resolve the links, just make
 sure that they are resolvable given the version of R and the package
 dependencies. 

 Robert


> 
> -- 
> 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
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

-- 
+---------------------------------------------------------------------------+
| Robert Gentleman                 phone : (617) 632-5250                   |
| Associate Professor              fax:   (617)  632-2444                   |
| Department of Biostatistics      office: M1B20                            |
| Harvard School of Public Health  email: rgentlem at jimmy.harvard.edu        |
+---------------------------------------------------------------------------+