I now reinstalled XCode and build the gcc-4.7 new, explicitly telling to build 
a objc compiler. Furthermore I installed the XQuartz package and can now use 
the symbolic link /usr/X11. 

1. (Configure) Running the command: 

../configure CC="gcc-4.7 -arch $arch" CXX="g++-4.7 -arch $arch" 
F77="gfortran-4.7 -arch $arch" FC="gfortran-4.7 -arch $arch" OBJC="gcc-4.7 
-arch $arch" --x-includes=/usr/X11/include/ --x-libraries=/usr/X11/lib/ 
--with-blas='-framework libVec' --with-lapack

or

../configure CC="gcc-4.7 -arch $arch" CXX="g++-4.7 -arch $arch" 
F77="gfortran-4.7 -arch $arch" FC="gfortran-4.7 -arch $arch" OBJC="objc-4.7 
-arch $arch" --x-includes=/usr/X11/include/ --x-libraries=/usr/X11/lib/ 
--with-blas='-framework libVec' --with-lapack

Gives no errors nor warnings in the configure step.

2. (Make) Doing a make with:

make -j 8

gives me still errors with aqua: http://pastebin.com/n3g6zYfS

If I try to build without aqua I do the following:

1. (Configure) I configure with: 

../configure CC="gcc-4.7 -arch $arch" CXX="g++-4.7 -arch $arch" 
F77="gfortran-4.7 -arch $arch" FC="gfortran-4.7 -arch $arch" OBJC="objc-4.7 
-arch $arch" --x-includes=/usr/X11/include/ --x-libraries=/usr/X11/lib/ 
--with-blas='-framework libVec' --with-lapack --without-aqua

All runs fins - no errors nor warnings. 

2. (Make) Running the command::

make -j 8

gives again errors with X11 (but quite different): http://pastebin.com/HXqLgH4b

Is there still a possible way to build R from sources with gcc-4.7?


Best regards

Simon

Begin forwarded message:

> From: Prof Brian Ripley <[email protected]>
> Subject: Re: [R-SIG-Mac] R compile from sources: error in make with gcc-4.7
> Date: September 13, 2012 10:33:32 PM GMT+02:00
> To: [email protected]
> 
> On 13/09/2012 20:48, Simon Urbanek wrote:
>> It seems your compiler doesn't support ObjC. You should get a working
>> compiler... Besides, gcc-4.7.x has been reported to break so your
>> mileage may vary even if you get Apple driver and a more complete
>> compiler. More comments inline below.
>> 
>> Cheers, Simon
>> 
>> 
>> On Sep 13, 2012, at 3:21 PM, Simon Zehnder wrote:
>> 
>>> Hi R-fellows,
>>> 
>>> I have some fatal errors compiling the R source files with
>>> gcc-4.7.1 on my Mountain Lion and I wonder if this is only on my
>>> Mac or if it does happen on other Macs too (I use the gcc-4.7 as I
>>> want to compile some C++ code that uses OpenMP 3.1).
>>> 
>>> Here is the way I did it:
>>> 
>>> 1. (gcc-4.7) Install gcc-4.7 on my Mac following
>>> http://solarianprogrammer.com/2012/07/21/compiling-gcc-4-7-1-mac-osx-lion/
>>> (This runs quite well)
>>> 
>>> 2. (Configure) Getting the R-2.15.1.tar.gz from
>>> http://cran.r-project.org,
>>> 
>>> I used on the shell:
>>> 
>>> tar -zxvf R-2.15.1.tar.gz cd R-2.15.1* mkdir build && cd build
>>> 
>>> Further, for the R configure to work it needs the XCode Command
>>> Line Tools
>>> 
>>> Now as the gcc-4.7 is not the default compiler on my Mac and R
>>> needs X11 the configure line is a little nasty but works without
>>> errors (see also
>>> http://stat.ethz.ch/R-manual/R-patched/doc/manual/R-admin.html#Installing-R-under-_0028Mac_0029-OS-X):
>>> 
>>> 
>>> 
> ../configure CC="gcc-4.7 -arch x86_64" CXX="g++-4.7 -arch x86_64" 
> F77="gfortran-4.7 -arch x86_64" FC="gfortran-4.7 -arch x86_64" OBJC="gcc-4.7 
> -arch x86_64" 
> -x-libraries=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.7.sdk/usr/X11/lib
>  
> -x-includes=/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.7.sdk/usr/X11/include
>  --with-lapack --with-blas='framework vecLib'
>>> 
>>> (The usual installation given on
>>> http://cran.r-project.org/bin/macosx/RMacOSX-FAQ.html, i.e.:
>>> 
>>> ../configure CC="gcc-4.7 -arch $arch" CXX="g++-4.7 -arch $arch"
>>> F77="gfortran-4.7 -arch $arch" FC="gfortran-4.7 -arch $arch"
>>> OBJC="gcc-4.7 -arch $arch" --x-includes=/usr/X11/include
>>> --x-libraries=/usr/X11/lib --with-blas='framework vecLib'
>>> --with-lapack
>>> 
>>> having defined arch=x86_64
>>> 
>>> does not work on my machine, as it seems that XCode installs the
>>> Command Line Tools with the X11SDK in
>>> /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX10.7.sdk/usr/X11.
>> 
>>> 
>> Again, possibly a problem with your compiler, a supported compiler
>> doesn't need that hack. But also 10.8 doesn't come with X11 so I
>> assume you installed XQuartz accordingly ...
>> 
>> 
>>> Compiling with this line usually gives:
>>> 
>>> configure: error: --with-x=yes (default) and X11 headers/libs are
>>> not available)
>>> 
>> 
>> Note that the instructions are for Apple compilers and those
>> compatible (e.g. FSF builds with Apple driver driver which adds extra
>> flags and parses others). For any other compilers, you're on your own
>> and the flags may vary.
>> 
>> 
>>> 3. (Make) Doing now the make gives me huge amount of fatal errors
>>> with no clue where it comes from and no google documentations so
>>> far. All make output can be seen here:
>>> http://pastebin.com/3baL8php
>>> 
>>> I have really no clue where these come from,
>> 
>> They some from including AppKit, apparently your compiler doesn't
>> even work with system frameworks. Even if you fix this instance there
>> is Quartz as well. There are a lot of broken compiler builds around,
>> so I'm not surprised. If you have to avoid all system services, you
>> can try to build without aqua, but I don't think that has been tested
>> recently on OS X ...
> 
> Would yesterday do?
> 
>> 
>> 
>>> but maybe somebody of you guys has encountered either these errors
>>> as well or has found another way to compile R with gcc-4.7?
>>> 
>>> 
>>> Best Regards
>>> 
>>> Simon
>>> 
>>> [[alternative HTML version deleted]]
>>> 
>>> _______________________________________________ R-SIG-Mac mailing
>>> list [email protected]
>>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>>> 
>>> 
>> 
>> _______________________________________________ R-SIG-Mac mailing
>> list [email protected]
>> https://stat.ethz.ch/mailman/listinfo/r-sig-mac
>> 
> 
> 
> -- 
> Brian D. Ripley,                  [email protected]
> 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
> 
> _______________________________________________
> R-SIG-Mac mailing list
> [email protected]
> https://stat.ethz.ch/mailman/listinfo/r-sig-mac


        [[alternative HTML version deleted]]

_______________________________________________
R-SIG-Mac mailing list
[email protected]
https://stat.ethz.ch/mailman/listinfo/r-sig-mac

Reply via email to