[R-SIG-Mac] Error in gfortran packaged with R binaries

[stefano iacus]

Bill,
it seems to me that Simon discourage to use fortran from (old) R  
binary and, indeed 2.4.1 boundle does not contain fortran compiler as  
far as I remember (I cannot check right now).
Also, gfortran from hpc.sourge* is the way to go.

Simon is actually in NZ so maybe we need some time to get some answer.
My current setup is Apple Gcc + gfrotran from hpc.*



On 14/feb/07, at 04:39, Bill Northcott wrote:

> I am building R from source using Apple's C/C++ compilers from Xcode
> 2.4.1 and the gfortran included in recent binary distributions.  The
> host is a Core Duo MacBook Pro.
>
> During the build I get a lot of these:
> /usr/local/gcc4.0/bin/gfortran -D__NO_MATH_INLINES -fPIC  -arch i386
> -c ../../../../src/extra/blas/cmplxblas.f -o cmplxblas.o
> /var/tmp//ccbSWVl5.s:23289:indirect jmp without `*'
> /var/tmp//ccbSWVl5.s:23304:indirect jmp without `*'
>
> which seems to indicate errors in the assembler output by the fortran
> compiler.
>
> This does not happen if I use a recent gfortran from
> hpc.sourceforge.net.  It does not happen with ppc targets using the R
> distributed compiler.  It DOES happen with an i386 target build on a
> ppc host.
>
> Bill Northcott
>
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac