Our lab has a lot of different unix boxes, with different hardware, and
I'm assuming (perhaps wrongly) that by setting a per-user package
installation directory, the packages will only work on one type of
hardware. Our systems are all set up to share the same home directory
(and, thus, the
The script .Rprofile evaluates R code on startup. You could use that
to test for various environment variables. Alternatively, use Unix
shell scripts to set system environment variables to be used in a
generic .Renviron. See help(Startup) for more details.
/Henrik
On Sun, Jan 25, 2009 at
On Sun, 25 Jan 2009, Henrik Bengtsson wrote:
The script .Rprofile evaluates R code on startup. You could use that
to test for various environment variables. Alternatively, use Unix
shell scripts to set system environment variables to be used in a
generic .Renviron. See help(Startup) for more
Ah, perfect -- so would the ideal R_LIBS_USER setting (to more or less
guarantee the libraries will work on every possible computer) be
something along the lines of:
~/myRlibraries/%V%p%o%a
Or is this overkill?
--j
Prof Brian Ripley wrote:
On Sun, 25 Jan 2009, Henrik Bengtsson wrote:
The
Jonathan Greenberg wrote:
Our lab has a lot of different unix boxes, with different hardware, and
I'm assuming (perhaps wrongly) that by setting a per-user package
installation directory, the packages will only work on one type of
hardware. Our systems are all set up to share the same home
On Sun, 25 Jan 2009, Jonathan Greenberg wrote:
Ah, perfect -- so would the ideal R_LIBS_USER setting (to more or less
guarantee the libraries will work on every possible computer) be something
along the lines of:
~/myRlibraries/%V%p%o%a
Or is this overkill?
%V is overkill. On some OSes %v
6 matches
Mail list logo