[R-SIG-Mac] udunits and RNetCDF ... simple or hard to get Mac version?

Simon Urbanek simon.urbanek at r-project.org
Fri Jan 15 15:32:59 CET 2010


Bob,

On Jan 14, 2010, at 8:08 PM, Robert Chatfield wrote:

> We have been using both ncdf and  RNetCDF
> packages, and often enjoy the general approach
> offered by RNetCDF
> 
> That package has not been available on the Mac for
> some time, it appears, ... but was maybe available
> last year  (see message below).
> 

AFAIR the problem with udunits (the library) is that it is not self-contained as it requires external files to be present. That means that the udunits package will not work properly unless you have those files. That is unfortunate since it prevents us from shipping packages based on it. We have in the past provided udunits without those external files thus not fully functional (albeit for the tests we did have the files installed), but not recently. We could do that again if desired but it is not very clean ...


> According to the CRAN documentation page and my own
> attempts to acquire RNetCDF,  the problem
> appears to be with the needed package "udunits"
> 
> That package is also useful for atmosphere, oceans,
> and connected sciences.
> 
> With a MacPro 1.1 running 10.5.8 and the indicated R and RGui,
> I get this result
> 
> * installing *source* package Œudunits‚ ...
> creating cache ./config.cache
> checking how to run the C preprocessor... gcc -arch i386 -std=gnu99 -E
> checking for gcc... gcc -arch i386 -std=gnu99
> checking whether the C compiler (gcc -arch i386 -std=gnu99 -g -O2 )  
> works... no
> configure: error: installation or configuration problem: C compiler  
> cannot create executables.

As Dave said this is likely a problem in your setup. Make sure you have recent Xcode installed and check config.log for the exact error (i.e. unpack udunits and run R CMD INSTALL udunits -- then you'll find config.log inside).

Cheers,
Simon



> ERROR: configuration failed for package Œudunits‚
> * removing Œ/Library/Frameworks/R.framework/Versions/2.10/Resources/ 
> library/udunits‚
> 
> The downloaded packages are in
> 	Œ/private/var/folders/pb/pbT1eFNa2RaEvE+1Yx812U+++TI/-Tmp-/RtmpzhkbgP/ 
> downloaded_packages‚
>> 
>> 
>> sessionInfo()
> R version 2.10.0 (2009-10-26)
> i386-apple-darwin9.8.0
> 
> locale:
> [1] en_US.UTF-8/en_US.UTF-8/C/C/en_US.UTF-8/en_US.UTF-8
> 
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> 
> loaded via a namespace (and not attached):
> [1] tools_2.10.0
> Warning message:
> In install.packages(c("udunits"), lib = "/Library/Frameworks/ 
> R.framework/Resources/library/",  :
>   installation of package 'udunits' had non-zero exit status
>> 
> 
> ***************
> Is this hard to fix?
> Could there be a RNetCDF without the udunits package?  .... I myself
> use udunits rather rarely.
> 
> Bob Chatfield
> 
> 
> 
> 
>>> Message: 2
>>> Date: Sat, 4 Jul 2009 14:35:40 -0700
>>> From: Jan de Leeuw <deleeuw at stat.ucla.edu>
>>> Subject: Re: [R-SIG-Mac] Unable to install package RGtk2
>>> To: "r-sig-mac at stat.math.ethz.ch Mac List"
>>> 	<r-sig-mac at stat.math.ethz.ch>
>>> Cc: statcompute <statcompute at mta.stat.ucla.edu>
>>> Message-ID: <FD535A21-B873-4986-8466-876455B77C1C at stat.ucla.edu>
>>> Content-Type: text/plain; charset=windows-1252; format=flowed;
>>> 	delsp=yes
>>> 
>>> I have a 2.5 GB file localxxxxxx.tgz, with xxxxxx indicating the  
>>> date.
>>> 
>>> You can get it from afp://gifi.stat.ucla.edu and then
>>> from my Public directory in /software. Or from a WebDAV connection to
>>> http://public.me.com/jdeleeuw/software (it's currently gone from
>>> me.com, but
>>> it will be there again soon).
>>> 
>>> It contains
>>> 
>>> a. 64-bit version of R-devel (a.k.a R-2.10.0) with about 600  
>>> packages.
>>> It supports the
>>>   GTK+, tcl/tk, SciViews, etc interfaces.
>>> 
>>> b. 64-bit version of grass-trunk (a.k.a. GRASS-7.0).
>>> 
>>> c. Support-libraries and executables, all 64-bit, for ImageMagick,
>>> freetype, GDAL, proj,
>>>   gsl, gmp, mpfr, fftw, mysql, pgsql, hadoop, root, heasoft, hdf4/5,
>>> netcdf, graphviz,
>>>   ffmpeg, asymptote, expat, cdd, cfitsio, fontconfig, gc, geos,
>>> geotiff, gd, tiff, jpeg,
>>>   png, svg, gettext, harminv, iconv, intl, jags, jasper, las,
>>> pixman, qhull, readline,
>>>   sqlite, szip, tcl/tk, udunits, xml2, gtk2, gdk, atk, glib, cairo,
>>> pango, glade, git,
>>>   wget, gawk, sed, gs, pkg-config, tramo, seats, x12a, yacas, pvm3
>>> --- probably more.
>>> 
>>> d. A 64-bit version of gcc, g++, gfortran 4.2.5 with gomp.  
>>> Everything,
>>> including the
>>>   compilers, is compiled with flags "-m64 -O3 -fopenmp -
>>> mtune=native" -- "native" in
>>>   this context means "nocona", i.e. Intel Core Duo.
>>> 
>>> e. texlive-2008
>>> 
>>> It is updated one or twice per week with recent cvs, svn or git
>>> versions. It installs in
>>> /usr/local. The R is pretty snappy on my machine.
>>> 
>>> !!!!!!!!!!!!!!!!!
>>> Limitations 
>>> !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
>>> 
>>> 1. This is for CLI junkies. Normal people will stay away from this.
>>> 
>>> 2. The quartz device in R works, but many other things are either GTK
>>> +, tcl/tk, or even
>>>   Athena.
>>> 
>>> 3. This was compiled on OS X 10.6 (Snow Leopard). Your Leopard  
>>> mileage
>>> may vary. No
>>>   universal binaries of any sort, of course.
>>> 
>>> 4. Basically it creates a separate 64-bit development environment,
>>> independent of the Apple stuff
>>>   (and independently, if you want, of the "normal" and nice R and
>>> GRASS framework-builds and GUIs).
>>> 
>>> 5. This is a large file. After unpacking "du -s /usr/local" says
>>> 14167872 (i.e. 14 GB).
>>> 
>>> 6. I am not really supporting this for users outside my department.
>>> That would be unwise.
> 
> 
> Dr. Robert Chatfield
> Earth Sciences, MS 245-5
> NASA Ames Research Center
> Moffett Field, CA 94035    USA
> 
> Ph: 650-604-5490  FAX 650-604-3625
> 
> http://geo.arc.nasa.gov/sgg/chatfield
> 
> 
> 
> 
> 
> 	[[alternative HTML version deleted]]
> 
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac



More information about the R-SIG-Mac mailing list