[R-sig-Geo] Problem installing rgeos
Waichler, Scott R
Scott.Waichler at pnnl.gov
Thu Aug 6 19:57:29 CEST 2015
First, on my system I reinstalled and ran ldconfig for the following software: geos-2.4.2, proj-4.9.1 and gdal-2.0.0. Within R I reinstalled Rcpp from source.
Along the way I had to address other problems. I deleted an empty directory called gdalplugins, and created a file /etc/ld.so.conf.d/gdal.conf with the line "/usr/local/lib" in it. I could finally install rgdal with
> install.packages("rgdal", dependencies = F, repos = "http://cran.fhcrc.org/", configure.args=c(rgdal='--with-gdal-lib=/usr/local/lib'))
and after that, I was also able to install rgeos. I'm not implying that rgeos depends on rgdal, but the command
R CMD check rgeos_0.3-11.tar.gz
brought up dependency errors that forced me to go back and get rgdal working first. I may be forgetting some key steps that made this all work, if so, my apologies. I hope I don't have to do it again!
> The problem is almost certainly that the platform build train (compilers and
> libraries) which underlie all the other programs have been updated, as Chris
> has suggested (and we still do not know the OS version or output of
> sessionInfo()). This is neither a GEOS nor an rgeos problem. Most likely the
> C and C++ libraries have changed, leading to a need to re-install an unknown
> number of downstream components.
> I suggest checking that you can install Rcpp from source (to test the build
> train). If there are problems, you should correct these first, by bringing your
> build train into reasonable order.
> If you can, re-install GEOS from source (to freshen up its dependencies on
> the C and C++ libraries) and only when GEOS passes make check itself
> (ourside R), proceed to install rgeos. If you download the R source tarball of
> rgeos, and run R CMD check on it, you will get a full test of its
> state. Following R CMD INSTALL, this should then work.
> > HTH
> > Chris
> > On Mon, Aug 3, 2015 at 10:34 AM, Waichler, Scott R
> > <Scott.Waichler at pnnl.gov>
> > wrote:
> >> Thanks for the suggestions, but I think what I need is help tracking
> >> down what the exact error is in the compilation. Let me try again
> >> with the part of the output describing the error:
> >>> install.packages("rgeos", dependencies = F, repos = "
> >> http://cran.fhcrc.org/")
> >> . . .
> >> checking for C compiler default output file name... a.out checking
> >> for suffix of executables...
> >> checking whether we are cross compiling... configure: error: in
> >> `/tmp/RtmpUjYUle/R.INSTALL41c15690b9e1/rgeos':
> >> configure: error: cannot run C compiled programs.
> >> If you meant to cross compile, use `--host'.
> >> See `config.log' for more details
> >> ERROR: configuration failed for package 'rgeos'
> >> * removing '/files3/R/R-3.2.1_install/lib64/R/library/rgeos'
> >> It says I should refer to config.log for more details. The only
> >> config.log file I can find in the directory tree is the one for geos,
> >> which I installed a couple of years ago, and did not give me problems
> >> with previous installations of rgeos under earlier releases of R. In
> >> case there is a clue, here is the output in that config.log around
> >> the cross compile
> >> message:
> >> configure:3682: result:
> >> configure:3704: checking whether we are cross compiling
> >> configure:3712: gcc -o conftest conftest.c >&5
> >> configure:3716: $? = 0
> >> configure:3723: ./conftest
> >> Is there a config.log for the attempted rgeos installation? Where
> >> would I find it?
> >>> You're build is failing pretty early in the configuration process
> >>> which suggests you are missing some tools in your development
> >>> environment. The latest geos was found. Was that a binary or built
> >>> from source? If from source you have the necessary tools in your
> >>> dev environment and it may merely be a function of dropping the
> >>> depends in you install.packages call....
> >> Rgeos was built from source.
> >> Thanks,
> >> Scott Waichler
> >> PNNL
> > [[alternative HTML version deleted]]
> > _______________________________________________
> > R-sig-Geo mailing list
> > R-sig-Geo at r-project.org
> > https://stat.ethz.ch/mailman/listinfo/r-sig-geo
> Roger Bivand
> Department of Economics, Norwegian School of Economics, Helleveien 30,
> N-5045 Bergen, Norway.
> voice: +47 55 95 93 55; fax +47 55 95 91 00
> e-mail: Roger.Bivand at nhh.no
More information about the R-sig-Geo