p_connolly at slingshot.co.nz
2009-Jul-09 01:45 UTC
[R] Mysteriously vanishing LD_LIBRARY_PATH
Using R-2.8.0 and R-2.8.1, I get behaviour like this: R version 2.8.0 (2008-10-20) Copyright (C) 2008 The R Foundation for Statistical Computing ISBN 3-900051-07-0 [....]> Sys.getenv("LD_LIBRARY_PATH")LD_LIBRARY_PATH "/home/hrapgc/local/R-2.8.0/lib:/usr/local/lib:/usr/lib/gcj-4.1.1:/home/hrapgc/Rstuff/library/asreml/libs"> q()Save workspace image? [y/n/c]: n However: R version 2.9.1 (2009-06-26) Copyright (C) 2009 The R Foundation for Statistical Computing ISBN 3-900051-07-0 [...]> Sys.getenv("LD_LIBRARY_PATH")LD_LIBRARY_PATH "/home/hrapgc/local/R-2.9.1/lib:/usr/local/lib:/usr/lib/gcj-4.1.2" I'm particularly dismayed by the fact that there are now only three parts to LD_LIBRARY_PATH, and slightly mystified by the appearance of /usr/lib/gcj-4.1.2 in place of /usr/lib/gcj-4.1.1. Maybe that's a red herring since the former is a link to the latter. I would appreciate help getting the rest of LD_LIBRARY_PATH back. I set it in ~/.bash_profile in case that makes a difference to the possible cause of the phenomenon. $ uname -r 2.6.18-92.1.22.el5 The problem doesn't seem to arise with Fedora 8 nor with Mepis 6.5. TIA -- Patrick Connolly HortResearch Mt Albert Auckland New Zealand Ph: +64-9 925 7079 ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~ I have the world`s largest collection of seashells. I keep it on all the beaches of the world ... Perhaps you`ve seen it. ---Steven Wright ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
R respects LD_LIBRARY_PATH, see /usr/lib/R/etc/ldpaths where it
prepends its own path to any value in the environment when R is
invoked:
if test -z "${LD_LIBRARY_PATH}"; then
LD_LIBRARY_PATH="${R_LD_LIBRARY_PATH}"
else
LD_LIBRARY_PATH="${R_LD_LIBRARY_PATH}:${LD_LIBRARY_PATH}"
fi
export LD_LIBRARY_PATH
====
Provide the output of
echo $LD_LIBRARY_PATH
or
env | grep LD_LIBRARY_PATH
to check that it's set correctly *BEFORE* you invoke R.
- Godmar
On Wed, Jul 8, 2009 at 9:45 PM, <p_connolly at slingshot.co.nz>
wrote:> Using R-2.8.0 and R-2.8.1, I get behaviour like this:
>
> R version 2.8.0 (2008-10-20)
> Copyright (C) 2008 The R Foundation for Statistical Computing
> ISBN 3-900051-07-0
>
> [....]
>
>> Sys.getenv("LD_LIBRARY_PATH")
>
> ? ? ? ? ?LD_LIBRARY_PATH
>
"/home/hrapgc/local/R-2.8.0/lib:/usr/local/lib:/usr/lib/gcj-4.1.1:/home/hrapgc/Rstuff/library/asreml/libs"
>> q()
> Save workspace image? [y/n/c]: n
>
>
> However:
>
> R version 2.9.1 (2009-06-26)
> Copyright (C) 2009 The R Foundation for Statistical Computing
> ISBN 3-900051-07-0
>
> [...]
>
>> Sys.getenv("LD_LIBRARY_PATH")
> ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? LD_LIBRARY_PATH
>
"/home/hrapgc/local/R-2.9.1/lib:/usr/local/lib:/usr/lib/gcj-4.1.2"
>
>
> I'm particularly dismayed by the fact that there are now only three
> parts to LD_LIBRARY_PATH, and slightly mystified by the appearance of
> /usr/lib/gcj-4.1.2 in place of /usr/lib/gcj-4.1.1. ?Maybe that's a red
> herring since the former is a link to the latter.
>
> I would appreciate help getting the rest of LD_LIBRARY_PATH back. ?I
> set it in ~/.bash_profile in case that makes a difference to the
> possible cause of the phenomenon.
>
> $ uname -r
> 2.6.18-92.1.22.el5
>
> The problem doesn't seem to arise with Fedora 8 nor with Mepis 6.5.
>
> TIA
>
> --
> Patrick Connolly
> HortResearch
> Mt Albert
> Auckland
> New Zealand
> Ph: +64-9 925 7079
> ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
> I have the world`s largest collection of seashells. I keep it on all
> the beaches of the world ... Perhaps you`ve seen it. ?---Steven Wright
> ~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~.~
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
On Jul 8, 2009, at 8:45 PM, p_connolly at slingshot.co.nz wrote:> Using R-2.8.0 and R-2.8.1, I get behaviour like this: > > R version 2.8.0 (2008-10-20) > Copyright (C) 2008 The R Foundation for Statistical Computing > ISBN 3-900051-07-0 > > [....] > >> Sys.getenv("LD_LIBRARY_PATH") > > LD_LIBRARY_PATH > "/home/hrapgc/local/R-2.8.0/lib:/usr/local/lib:/usr/lib/gcj-4.1.1:/ > home/hrapgc/Rstuff/library/asreml/libs" >> q() > Save workspace image? [y/n/c]: n > > > However: > > R version 2.9.1 (2009-06-26) > Copyright (C) 2009 The R Foundation for Statistical Computing > ISBN 3-900051-07-0 > > [...] > >> Sys.getenv("LD_LIBRARY_PATH") > LD_LIBRARY_PATH > "/home/hrapgc/local/R-2.9.1/lib:/usr/local/lib:/usr/lib/gcj-4.1.2" > > > I'm particularly dismayed by the fact that there are now only three > parts to LD_LIBRARY_PATH, and slightly mystified by the appearance of > /usr/lib/gcj-4.1.2 in place of /usr/lib/gcj-4.1.1. Maybe that's a red > herring since the former is a link to the latter. > > I would appreciate help getting the rest of LD_LIBRARY_PATH back. I > set it in ~/.bash_profile in case that makes a difference to the > possible cause of the phenomenon. > > $ uname -r > 2.6.18-92.1.22.el5 > > The problem doesn't seem to arise with Fedora 8 nor with Mepis 6.5. > > TIAPatrick, When I was using Fedora, several years ago I started to add any paths that I needed for LD_LIBRARY_PATH into /etc/ld.so.conf. Then run 'sudo ldconfig' to update the configuration. I got bit by the same issue that you are seeing when using RODBC with Oracle. I believe that it was Prof. Ripley who suggested that I use the above approach and did not have any problems after that. HTH, Marc Schwartz