[R] R: error while loading shared libraries: libg2c.so.o

Prof Brian Ripley ripley at stats.ox.ac.uk
Wed Jan 19 23:34:23 CET 2005


On Wed, 19 Jan 2005, Peter Dalgaard wrote:

> "Jon Dressel" <jdressel at surromed.com> writes:
>
>>> If you installed an RPM, please tell us so, and which one from where
>>
>> The RPM I installed is R-2.0.1-0.RH3AS.i386.rpm from Index of 
>> /bin/linux/redhat/el3/i386 located at http://cran.stat.ucla.edu/ .
>>
>>> I very much suspect it is libg2c.so.0.
>>> It should be in /usr/lib if you have the prerequisites installed.
>>
>> It is currently installed in /usr/lib64
>
> So it's an x86_64 system? If so, what do you want an i386 RPM of R
> for? You could well be better off compiling from source. If you
> insist, then you might get it to work by installing the 32bit version
> of the library in /usr/lib, but there's no guarantee, and you probably
> get into further trouble the first time you try installing a CRAN
> package.
>
>>> You need to know where R's home is.  Normally R RHOME will tell you, but
>>> that might give the same error.  So use
>>
>>> grep '^R_HOME_DIR' `which R`
>>
>>> and then substitute it in
>>
>>> R CMD ldd R_HOME_DIR/bin/exec/R
>
>
>> I did the above and determined that R_HOME_DIR is /usr/lib/R . I
>> plugged in above and did not get any of the description as in your
>> example and received the same error message:
>
> I think that was a typo; just "ldd /usr/lib/R/bin/exec/R".

It was deliberate and correct, as R CMD sets the LD_LIBRARY_PATH.
ldd /usr/lib/R/bin/exec/R gives unsatisfied references on many systems.

I just do not believe references to libg2c.so.o, though.
                                               ^
If that is really what is being seen, that RPM is corrupt.

-- 
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




More information about the R-help mailing list