[Rd] Installed file locations (PR#1118)

ripley@stats.ox.ac.uk ripley@stats.ox.ac.uk
Sat, 6 Oct 2001 16:21:27 +0200 (MET DST)

On Fri, 5 Oct 2001 gml4410@ggr.co.uk wrote:

>    R (1.3.1) appears to have a bug (well a mis-feature, I suppose).

I believe you mis-understand the wisdom of the feature.

>    It places architecture-specific and architecture-dependent files in
> the same locations.  This means that the common files cannot be shared
> across installations on multiple architectures.
>    It would be nice if it could separate these under exec_prefix and
> prefix and look in the relevant one for each file.

What would be `nice' about it? There are `architecture-specific' (also
known as `architecture-dependent', sic) files all over the place in R.
There are few files that are not platform-dependent: for example many of
the help files and the base R code file are.  Many of these are not
executables and do not belong under `exec_prefix'.

This cannot be a space issue this century (as it was originally I believe),
as surely you can afford the disc space for copies of R for each of a
sensible number of platforms?

Suppose one did combine the trees for i386-linux and ppc-linux.  What do
you think would happen when i386-linux is upgraded and ppc-linux is not?
And R is designed to be upgraded on the fly, using update.packages().  We
have public R installations for sparc-solaris, i386-linux and Windows and
sharing them for R would be a nightmare for the administrators.  (We do
share them for TeX, for example.)

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 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch