[R-SIG-Mac] "-fno-automatic" unrecognized by new gfortran

[Simon Urbanek]

Adam,

unfortunately gfortran invokes cc1 for your code (I'm not sure why),  
which comes from the "official" tools that don't support Fortran. For  
your particular setup this should be easily fixable by using cc1 from  
the build that included Fortran support:
http://r.research.att.com/cc1-gcc42fort5531-darwin9.tar.gz
but in general this poses a problem, because Apple's version of cc1  
doesn't know about the Fortran front-end (and I don't feel like  
messing with their binaries). Given that I'm using Apple's sources, it  
should be fine for now, but you never know.

Thanks for raising the issue.
Simon

On Nov 28, 2007, at 3:25 PM, Adam Kocoloski wrote:

> Hi,
>
> I tried the new gfortran-42.pkg and gcc42build5531.pkg hosted at
> r.research.att.com/tools (ADC Student membership isn't enough to get
> me the Developer Preview of gcc 4.2 from Apple), and I found that the
> "-fno-automatic" option isn't recognized:
>
> /usr/bin/gfortran -arch ppc -arch 'i386' -arch ppc64 -arch x86_64 -c -
> O0 -g -funroll-loops -fomit-frame-pointer -ftree-vectorize -fno-
> automatic   -I/Users/kocolosk/Documents/src/cernlib/2006/build/mclibs/
> cojets -I/Users/kocolosk/Documents/src/cernlib/2006/src/mclibs/ 
> cojets -
> I/Users/kocolosk/Documents/src/cernlib/2006/src/mclibs/cojets/code -I/
> Users/kocolosk/Documents/src/cernlib/2006/src/include  -DCERNLIB_LINUX
> -DCERNLIB_UNIX -DCERNLIB_LNX -DCERNLIB_QMGLIBC -DCERNLIB_MACOSX -
> DCERNLIB_GFORTRAN   -o archive/acosx.o /Users/kocolosk/Documents/src/
> cernlib/2006/src/mclibs/cojets/code/acosx.F
> cc1: error: unrecognized command line option "-fno-automatic"
> cc1: error: unrecognized command line option "-fno-automatic"
> cc1: error: unrecognized command line option "-fno-automatic"
> cc1: error: unrecognized command line option "-fno-automatic"
>
> I had previously used a gfortran built from FSF sources using the gcc-
> universal.sh script from the same site without problems.  Sorry if
> this isn't the appropriate place to file a bug report /
> troubleshooting request against those compilers.  Best,
>
> Adam
> -----------
> Adam Kocoloski
> Dept. of Physics, MIT
>
> _______________________________________________
> R-SIG-Mac mailing list
> R-SIG-Mac at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>
>