[BioC] RBGL 1.8.1 installation error on linux
rgentlem
rgentlem at fhcrc.org
Thu May 11 18:46:08 CEST 2006
Hi,
Johannes Rainer wrote:
> hi,
>
> thanks robert for your help.
> based on your suggestions i tested now some new settings, i used the gcc
> 3.2.3 compiler (g++32 ) to compile the package (unfortunately there is
Hmm,
I don't think that is the right set of numbers for a gcc compiler,
4.0.3 and 3.4.5 are the up to date versions are reported at:
http://gcc.gnu.org/.
I don't know why there would not be a version of all for FC5,
certainly there were discussions of it and some of the early work on FC5
was based on the 4.0 arm (4.1 is very new, released 2006-02-28). You
should be able to build these from source if you wanted to.
I remain convinced that these are likely compiler bugs and hence
someone at Redhat or gcc might care, we are not really able to do much
about them. Sorry we don't have such a beast here for any testing so I
cannot say much more that would be helpful.
best wishes
Robert
> no 4.0 version available for FC5) and get, on my system with R 2.3 and
> RBGL_1.8.1 the same error as before. but interestingly the installation
> of RBGL_1.6.0.tar.gz with the g++32 worked (but not with the g++ from
> gcc 4.1).
> well, i guess i will have to try on...
>
> thanks for your help
> jo
>
> On 5/11/06, *rgentlem* <rgentlem at fhcrc.org <mailto:rgentlem at fhcrc.org>>
> wrote:
>
> Hi,
>
> Johannes Rainer wrote:
> > ok now i tested some other things. on a system (FC4) with the gcc
> 4.0 there
> > is no problem installing the package. could it be that it does
> not work with
> > gcc4.1? (although it compiles the source files neatly, but after the
> > compilation it gets that error).
>
> Yes, it is entirely possible that such versions of gcc will fail on
> something as heavily templated as BGL. gcc tends to be somewhat unstable
> for versions near the bleeding edge.
>
> You do not need a copy of BGL, it is included in the sources for the
> package and in fact, having a copy around could cause some problems as I
> think we (Li) have to patch from time to time (right now I think a
> header file) and if you compile against the wrong one bad things are
> going to happen.
>
> I has nothing to do with 64bit, we build on that, but Suse not Redhat.
>
> > i edited also the Makeconf file located in /usr/lib64/R/etc and
> set the the
> > CXXFLAGS to CXXFLAGS = -O2 -g -m64. nothing worked (i only got
> less warnings
> > during the compilation).
> > actually i run out of ideas what i could try next...
>
> Use a version of gcc that is known to work? Your tests and ours have
> indicated that gcc 4.0.x does work. I am afraid this looks like a
> redhat/gcc problem and you might want to take it up with them.
>
> best wishes
> Robert
>
> ps Li, I am not seeing your responses could you please make sure that
> they are too the list, this is archived and many find it useful to be
> able to follow the whole thread
>
> >
> > jo
> >
> >
> > On 5/11/06, Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>
> <Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>> wrote:
> >>
> >> That's right version. boost lib is packaged in RBGL. That's
> not what I
> >> would worry about.
> >>
> >> Li
> >>
> >>> which version of the boost library do i need? i have boost 1.33.1
> >>>
> >>> jo
> >>>
> >>> On 5/11/06, Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>
> <Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>> wrote:
> >>>>
> >>>> Pretty much all the msgs are about templates, so I would look
> more into
> >>>> what your g++/STL and the likes on your machine.
> >>>>
> >>>> Li
> >>>>
> >>>>> hi Li
> >>>>>
> >>>>> thanks for your quick reply,
> >>>>> concerning the extra parameters, i have not defined anything,
> i just
> >>>>> downloaded the RBGL_1.8.1.tar.gz and run R CMD INSTALL on it,
> i did
> >>>> not
> >>>>> specify any extra parameter... don't know from where they are
> coming.
> >>>>> actually the R i installed is a precompiled one, i got it
> from one of
> >>>> the
> >>>>> fedora repositories, before i installed this one i compiled it
> >> myself,
> >>>> but
> >>>>> was also not able to install RBGL.
> >>>>>
> >>>>> R CMD INSTALL --configure-args='-g -O2' RBGL_1.8.1.tar.gz
> >>>>>
> >>>>> also resulted in an error (see the attachment).
> >>>>> the version of g++ (i suppose it is an x86_64 version, at
> least it
> >>>>> supports
> >>>>> m32 and m64...):
> >>>>>
> >>>>> [jo at bugblatterbeast ~]$ g++ --version
> >>>>> g++ (GCC) 4.1.0 20060304 (Red Hat 4.1.0-3 )
> >>>>> Copyright (C) 2006 Free Software Foundation, Inc.
> >>>>> This is free software; see the source for copying
> conditions. There
> >>>> is
> >>>> NO
> >>>>> warranty; not even for MERCHANTABILITY or FITNESS FOR A
> PARTICULAR
> >>>>> PURPOSE.
> >>>>>
> >>>>> i will now try to install it on an other machine running FC5,
> perhaps
> >>>>> there
> >>>>> is something wrong with my system...
> >>>>>
> >>>>> thanks
> >>>>> jo
> >>>>>
> >>>>>
> >>>>> On 5/11/06, Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>
> <Li.Long at isb-sib.ch <mailto:Li.Long at isb-sib.ch>> wrote:
> >>>>>>
> >>>>>> Hi,
> >>>>>>
> >>>>>> I'll be really interested in finding out what's happening in
> your
> >>>>>> settings.
> >>>>>> Could you provide additional info
> >>>>>> o. why/how did you decide to use the extra parameters when
> >> compiling?
> >>>>>> (such as protect-stack, and the like) If you just use "-g
> -O2",
> >> what
> >>>> do
> >>>>>> you get?
> >>>>>> o. what version of g++ do you have? is it for 64bit?
> >>>>>>
> >>>>>> It's my package, I'll try to help you figure it out.
> >>>>>>
> >>>>>> Thanks
> >>>>>>
> >>>>>> Li
> >>>>>>
> >>>>>>> hi,
> >>>>>>> i'm just moving from R 2.2 and BioC 1.7 to R 2.3 and BioC
> 1.8, but
> >>>> got
> >>>>>> a
> >>>>>>> problem when i tried to install the RBGL package. i
> included the
> >>>> whole
> >>>>>>> output of the R CMD INSTALL call in the attachment, below
> is just
> >>>> the
> >>>>>>> error
> >>>>>>> message that causes the installation failure:
> >>>>>>>
> >>>>>>>
> >>>>>>> boostIncl/boost/graph/depth_first_search.hpp:324: error: no
> >>>> matching
> >>>>>>> function for call to 'vertices(const
> std::vector<std::vector<long
> >>>>>> unsigned
> >>>>>>> int, std::allocator<long unsigned int> >,
> >>>>>> std::allocator<std::vector<long
> >>>>>>> unsigned int, std::allocator<long unsigned int> > > >&)'
> >>>>>>> make: *** [interfaces.o] Error 1
> >>>>>>> chmod: cannot access `/usr/lib64/R/library/RBGL/libs/*': No
> such
> >>>> file
> >>>>>> or
> >>>>>>> directory
> >>>>>>> ERROR: compilation failed for package 'RBGL'
> >>>>>>> ** Removing '/usr/lib64/R/library/RBGL'
> >>>>>>>
> >>>>>>> i don't know if that is a problem of my system (running fedora
> >> core
> >>>> 5
> >>>>>> on
> >>>>>> a
> >>>>>>> amd64) or not. has anyone experienced the same problems, or
> has
> >>>> anyone
> >>>>>> a
> >>>>>>> solution to this problem?
> >>>>>>>
> >>>>>>> thanks in advance,
> >>>>>>> jo
> >>>>>>> --
> >>>>>>> Johannes Rainer, Msc
> >>>>>>> Tyrolean Cancer Research Institute
> >>>>>>> Innrain 66, 6020 Innsbruck, Austria
> >>>>>>> Tel.: +43 512 570485 15
> >>>>>>> Email: johannes.rainer at tcri.at <mailto:johannes.rainer at tcri.at>
> >>>>>>> johannes.rainer at tugraz.at
> <mailto:johannes.rainer at tugraz.at>
> >>>>>>> _______________________________________________
> >>>>>>> Bioconductor mailing list
> >>>>>>> Bioconductor at stat.math.ethz.ch
> <mailto: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
> >>>>>>
> >>>>>
> >>>>> --
> >>>>> Johannes Rainer, Msc
> >>>>> Tyrolean Cancer Research Institute
> >>>>> Innrain 66, 6020 Innsbruck, Austria
> >>>>> Tel.: +43 512 570485 15
> >>>>> Email: johannes.rainer at tcri.at <mailto:johannes.rainer at tcri.at>
> >>>>> johannes.rainer at tugraz.at
> <mailto:johannes.rainer at tugraz.at>
> >>>>> _______________________________________________
> >>>>> Bioconductor mailing list
> >>>>> Bioconductor at stat.math.ethz.ch
> <mailto: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
> <http://news.gmane.org/gmane.science.biology.informatics.conductor>
> >>>>
> >>>
> >>> --
> >>> Johannes Rainer, Msc
> >>> Tyrolean Cancer Research Institute
> >>> Innrain 66, 6020 Innsbruck, Austria
> >>> Tel.: +43 512 570485 15
> >>> Email: johannes.rainer at tcri.at <mailto:johannes.rainer at tcri.at>
> >>> johannes.rainer at tugraz.at
> <mailto:johannes.rainer at tugraz.at>
> >>>
> >>
> >
> >
>
> --
> Robert Gentleman, PhD
> Program in Computational Biology
> Division of Public Health Sciences
> Fred Hutchinson Cancer Research Center
> 1100 Fairview Ave. N, M2-B876
> PO Box 19024
> Seattle, Washington 98109-1024
> 206-667-7700
> rgentlem at fhcrc.org <mailto:rgentlem at fhcrc.org>
>
>
>
>
> --
> Johannes Rainer, Msc
> Tyrolean Cancer Research Institute
> Innrain 66, 6020 Innsbruck, Austria
> Tel.: +43 512 570485 15
> Email: johannes.rainer at tcri.at <mailto:johannes.rainer at tcri.at>
> johannes.rainer at tugraz.at <mailto:johannes.rainer at tugraz.at>
--
Robert Gentleman, PhD
Program in Computational Biology
Division of Public Health Sciences
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N, M2-B876
PO Box 19024
Seattle, Washington 98109-1024
206-667-7700
rgentlem at fhcrc.org
More information about the Bioconductor
mailing list