[R] 2 "linux/R environment" questions
Oleg Sklyar
osklyar at ebi.ac.uk
Wed Oct 11 16:29:01 CEST 2006
Hi, my post about failing JGR on amd64 solved and here is how I got it
running, hopefully it helps as well.
Suppose your JDK (and you need JDK, not only JRE) is in
/usr/lib/jvm/java-1.5.0-sun (thus JRE in
/usr/lib/jvm/java-1.5.0-sun/jre). Set JAVA_HOME to point there to JDK.
Run R CMD javareconf (thanks Prof. Ripley). RE-install rJava, install
JGR. Start JGR - worked for me with sun-java binary deb of Ubuntu Edgy
and R2.4.0alpha from source.
Best,
Oleg
Prof Brian Ripley wrote:
> On Wed, 11 Oct 2006, Stefan Grosse wrote:
>
>> Joe Byers schrieb:
>>> JGR, when I manually compiled it for my system by specifying my location
>>> of java 1.5.0_06, worked fairly nicely. The autoinstall does not find
>>> my installation of this java version since Redhat EL4 does not support
>>> it. Redhat EL5 will in the near future. Maybe JGR will autoinstall then.
>>>
>> Can you tell how you worked that out (the manual compilation)? I failed
>> on Fedora Core 5 with JRE 1.5.0_08 with install.packages and dep=true. I
>> tried a hint from the JGR mailing list but failed...
>
> Which arch and how did you install R? For me, it works with R compiled
> from the sources (but not installed from RPM) on FC5 i686, but not x86_64.
> The issue is a run-time one:
>
>> library(JGR)
> Loading required package: rJava
> Loading required package: JavaGD
> Loading required package: iplots
> createObject.FindClass org/rosuda/iplots/Framework failed
> Exception in thread "main" java.lang.InternalError
>
> and that class of problem seems all too common on amd64 versions of Sun
> JRE. (On FC3, I managed to make it work with a self-compiled gcc 4.1.1,
> and haven't yet tried that on FC5.)
>
> The key is to use alternatives(8) to make sure the right jre is found, or
> to set JAVA_HOME.
>
> If you install R from RPMs you will probably need to run R CMD javareconf
> to get the right environment variables set in R itself.
>
--
Dr Oleg Sklyar * EBI/EMBL, Cambridge CB10 1SD, England * +44-1223-494466
More information about the R-help
mailing list