[Rd] R-1.9.0 won't compile on OSX (PR#6848)

Prof Brian Ripley ripley at stats.ox.ac.uk
Mon May 3 14:52:24 CEST 2004

I don't think this is a bug in *R*, especially as you appear to be using 
the unreleased gcc 3.4!

Hint: there needs to be a dynamic library for libg2c, not a static one. If
there is not, that's a g77 on MacOS X problem, which at one time (gcc 3.1)
we had to overcome by using vecLib.

See the R-admin manual which says

  ./configure --with-blas='-framework vecLib' --with-lapack --with-aqua 

  The first two options are strongly recommended. 

Did you actually follow that advice?  You appear to be trying to use 
atlas, and there is advice about needing a PIC version for that to work.

On Mon, 3 May 2004 Dan.Kelley at Dal.Ca wrote:

> Full_Name: Dan Kelley
> Version: 1.9.0
> OS: OSX (Darwin 7.3.0)
> Submission from: (NULL) (
> R-1.9.0 will not compile on Apple OSX (Darwin kernel 7.3.0).  The error msg
> starts as follows, and has several more errors.
> gcc -dynamiclib -L/sw/lib -L/usr/local/lib -install_name
> /Library/Frameworks/R.f
> ramework/Versions/1.9.0/Resources/bin/libRlapack.dylib -o libRlapack.dylib
> dlapa
> ck0.lo dlapack1.lo dlapack2.lo dlapack3.lo cmplx.lo  -lf77blas -latlas
> -L/sw/lib
>  -L/usr/local/lib -L/sw/lib/gcc/powerpc-apple-darwin7.2.0/3.4
> -L/sw/lib/gcc/powe
> rpc-apple-darwin7.2.0/3.4/../../.. -lfrtbegin -lg2c -lSystem
> ld: common symbols not allowed with MH_DYLIB output format with the
> -multi_modul
> e option
> /sw/lib/libg2c.a(err.o) definition of common _f__cblank (size 4)
> /sw/lib/libg2c.a(fmt.o) definition of common _f__cnt (size 40)

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-devel mailing list