On Tue, 9 Feb 2010, Stephane Schicklin wrote: > Hi, > I'm not using your library but I think your problem is different to a version > problem. If your libSBML is correctly installed, I advise you to search on > your system where really is the shared object "rsbml.so", then you'll have to > correct the path environment of R to have a correct load. > ideas : is your system/R really 64 bits ? Look at > '/usr/lib*32*/R/library/rsbml/libs/rsbml.so' to see if your library is > compiled for 32 and not for 64 bits system... Hi Stephane, (sorry for the multiple emails, forgot the list in the first email) thanks, but no, i think everything is 64bit and other libs are lying at the same location as rsbml. I found a potential source of the problem, though. I have a local installation of libsbml which I use outside of R. Even though the local installation is in my PATH and LD_LIBRARY_PATH variables, biocLite('rsbml') didn't find the headers and I didn't find out how to pass the local libsbml path to the biocLite installation scripts (? btw). Thus i asked the sysadmin for a global libsbml installation. Installation of rsbml then worked fine, but ldd shows that the rsbml.so file then actually links to the local version: merlin ~> ldd /usr/lib64/R/library/rsbml/libs/rsbml.so ... libsbml.so => /MYLOCALLIBS/lib/libsbml.so ... So, it seems to be a version conflict, which I often had with libsbml local vs. global installations. i should have though of ldd'ing the libs first. I guess, I can try to remove the local installation temporarily from the path variables during rsbml installation to force it to use the global libsbml version, OR I could set the respective compilation flags via biocLite,if i find out how exactly (some tries with configure.args didn't work yet). I'll try tomorrow :) Thanks, Rainer > Best wishes, > > Stephane Schicklin > > > Le 09/02/10 16:16, Rainer Machne a ?crit : >> Hi, >> >> Which libSBML version is required for rsbml? I couldn't find this >> information anywhere, and it doesn't work for libSBML 3.4.1 (and 2.4.0) on >> Fedora Core 10 Linux. >> >> Or is there another problem? Installation ended without error message, but >> then loading the library fails: >> >> > library('rsbml',lib.loc="/usr/lib64/R/library/") >> Error in dyn.load(file, DLLpath = DLLpath, ...) : >> unable to load shared library '/usr/lib64/R/library/rsbml/libs/rsbml.so': >> /usr/lib64/R/library/rsbml/libs/rsbml.so: undefined symbol: >> SpeciesReferenceGlyph_getSpeciesReferenceId >> Error : .onLoad failed in 'loadNamespace' for 'rsbml' >> Error: package/namespace load failed for 'rsbml' >> >> Thanks, >> Rainer Machne >> >> _______________________________________________ >> Bioconductor mailing list >> Bioconductor at stat.math.ethz.ch >> https://stat.ethz.ch/mailman/listinfo/bioconductor >> Search the archives: >> http://news.gmane.org/gmane.science.biology.informatics.conductor >> > -- > St?phane Schicklin > Bioinformatics Engineer > > INSERM, U851 > Marvel's Team "Immunit?, Infection, Vaccination" > IFR128 (Biosciences Lyon Gerland) > 21, Av. Tony Garnier > 69007 LYON FRANCE > > Phone : + 33 (0)4 37 28 23 47 > Fax : + 33 (0)4 37 28 23 41 > Mail : stephane.schicklin at inserm.fr <mailto:stephane.schicklin at="" inserm.fr="">

On Tue, Feb 9, 2010 at 10:26 AM, Rainer Machne <raim@tbi.univie.ac.at>wrote: > > On Tue, 9 Feb 2010, Stephane Schicklin wrote: > > Hi, >> I'm not using your library but I think your problem is different to a >> version problem. If your libSBML is correctly installed, I advise you to >> search on your system where really is the shared object "rsbml.so", then >> you'll have to correct the path environment of R to have a correct load. >> ideas : is your system/R really 64 bits ? Look at >> '/usr/lib*32*/R/library/rsbml/libs/rsbml.so' to see if your library is >> compiled for 32 and not for 64 bits system... >> > > Hi Stephane, > (sorry for the multiple emails, forgot the list in the first email) > > thanks, but no, i think everything is 64bit and other libs are lying at the > same location as rsbml. I found a potential source of the problem, though. I > have a local installation of libsbml which I use outside of R. Even though > the local installation is in my PATH and LD_LIBRARY_PATH variables, > biocLite('rsbml') didn't find the headers and I didn't find out how to pass > the local libsbml path to the biocLite installation scripts (? btw). > > Thus i asked the sysadmin for a global libsbml installation. Installation > of rsbml then worked fine, but ldd shows that the rsbml.so file then > actually links to the local version: > > merlin ~> ldd /usr/lib64/R/library/rsbml/libs/rsbml.so > ... > libsbml.so => /MYLOCALLIBS/lib/libsbml.so > ... > > So, it seems to be a version conflict, which I often had with libsbml local > vs. global installations. i should have though of ldd'ing the libs first. > > I guess, I can try to remove the local installation temporarily from the > path variables during rsbml installation to force it to use the global > libsbml version, OR I could set the respective compilation flags via > biocLite,if i find out how exactly (some tries with configure.args didn't > work yet). I'll try tomorrow :) > > I think the latter solution would be best in the long term. The configure.args argument *should* work with biocLite, so it would be good if you pasted exactly what you were passing to the rsbml configure script. Thanks, Michael > Thanks, > Rainer > > > Best wishes, >> >> Stephane Schicklin >> >> >> Le 09/02/10 16:16, Rainer Machne a écrit : >> >> Hi, >>> >>> Which libSBML version is required for rsbml? I couldn't find this >>> information anywhere, and it doesn't work for libSBML 3.4.1 (and 2.4.0) on >>> Fedora Core 10 Linux. >>> >>> Or is there another problem? Installation ended without error message, >>> but then loading the library fails: >>> >>> > library('rsbml',lib.loc="/usr/lib64/R/library/") >>> Error in dyn.load(file, DLLpath = DLLpath, ...) : >>> unable to load shared library >>> '/usr/lib64/R/library/rsbml/libs/rsbml.so': >>> /usr/lib64/R/library/rsbml/libs/rsbml.so: undefined symbol: >>> SpeciesReferenceGlyph_getSpeciesReferenceId >>> Error : .onLoad failed in 'loadNamespace' for 'rsbml' >>> Error: package/namespace load failed for 'rsbml' >>> >>> Thanks, >>> Rainer Machne >>> >>> _______________________________________________ >>> Bioconductor mailing list >>> Bioconductor@stat.math.ethz.ch >>> https://stat.ethz.ch/mailman/listinfo/bioconductor >>> Search the archives: >>> http://news.gmane.org/gmane.science.biology.informatics.conductor >>> >>> -- >> Stéphane Schicklin >> Bioinformatics Engineer >> >> INSERM, U851 >> Marvel's Team "Immunité, Infection, Vaccination" >> IFR128 (Biosciences Lyon Gerland) >> 21, Av. Tony Garnier >> 69007 LYON FRANCE >> >> Phone : + 33 (0)4 37 28 23 47 >> Fax : + 33 (0)4 37 28 23 41 >> Mail : stephane.schicklin@inserm.fr <mailto:stephane.schicklin@inserm.fr> > > > _______________________________________________ > Bioconductor mailing list > Bioconductor@stat.math.ethz.ch > https://stat.ethz.ch/mailman/listinfo/bioconductor > Search the archives: > http://news.gmane.org/gmane.science.biology.informatics.conductor > [[alternative HTML version deleted]]