math/R fails to build

[Philipp Ost]

bf wrote:
> --- On Tue, 1/13/09, Glen Barber <glen.j.barber at gmail.com> wrote:
> 
> 
>>From: Glen Barber <glen.j.barber at gmail.com>
>>Subject: Re: math/R fails to build
>>To: "Philipp Ost" <pj at smo.de>
>>Cc: bf2006a at yahoo.com, ports at freebsd.org
>>Date: Tuesday, January 13, 2009, 8:03 PM
>>On Tue, Jan 13, 2009 at 5:10 PM, Philipp Ost
>><pj at smo.de> wrote:
>>
>>>Hi,
>>>
>>>math/R fails to build:
>>># make
>>>[...]
>>>cc -std=gnu99      -I. -I../../src/include
>>
>>-I../../src/include
>>
>>>-I/usr/local/include -DHAVE_CONFIG_H  -fpic  -O2
>>
>>-fno-strict-aliasing -pipe
>>
>>>-march=athlon-xp -c Rmain.c -o Rmain.o
>>>cc -std=gnu99 -export-dynamic -L/usr/local/lib -o
>>
>>R.bin Rmain.o -L../../lib
>>
>>>-lR
>>>/usr/local/lib/libblas.so.2: undefined reference to
>>>`_gfortran_runtime_error_at'
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R/work/R-2.8.1/src/main.
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R/work/R-2.8.1/src/main.
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R/work/R-2.8.1/src.
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R/work/R-2.8.1.
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R.
>>>*** Error code 1
>>>
>>>Stop in /usr/ports/math/R.
>>>#
>>>
>>>I run RELENG_7 as of 23 Dec. 2008 (on i386).
>>>
>>>Any hints? (Apart from rebuilding blas, which I did.
>>
>>It didn't solve the
>>
>>>problem)
>>>
>>
>>Is it a safe assumption that you have recently updated your
>>ports tree?

Yes

[...]

> There were recently some changes to
> the organization of the Fortran-related
> ports on FreeBSD.  The Fortran compiler
> was added to the list of runtime
> dependencies, as it should have been,
> and the default Fortran compiler was
> changed from gcc 4.2.x to 4.3.x.  As
> Glen suggested, your problem may be
> related.  Try updating your ports tree,
> and then rebuilding blas, lapack,
> and R, and see if it works.
[...]

gcc43 is installed, ports tree is updated (as of ~20:00 MET today). I 
did rebuild lapack and blas, then tried R again:
# make
[...]
cc -std=gnu99      -I. -I../../src/include -I../../src/include 
-I/usr/local/include -DHAVE_CONFIG_H  -fpic  -O2 -fno-strict-aliasing 
-pipe -march=athlon-xp -c Rmain.c -o Rmain.o
cc -std=gnu99 -export-dynamic -L/usr/local/lib -o R.bin Rmain.o 
-L../../lib -lR
/usr/bin/ld: warning: libgfortran.so.2, needed by 
/usr/local/lib/libR.so, not found (try using -rpath or -rpath-link)
/usr/local/lib/libblas.so.2: undefined reference to `_gfortran_st_write'
/usr/local/lib/libblas.so.2: undefined reference to 
`_gfortran_transfer_character'
/usr/local/lib/libblas.so.2: undefined reference to 
`_gfortran_st_write_done'
/usr/local/lib/libblas.so.2: undefined reference to 
`_gfortran_runtime_error_at'
/usr/local/lib/libblas.so.2: undefined reference to 
`_gfortran_transfer_integer'
/usr/local/lib/libblas.so.2: undefined reference to `_gfortran_stop_numeric'
*** Error code 1

Stop in /usr/ports/math/R/work/R-2.8.1/src/main.
*** Error code 1

Stop in /usr/ports/math/R/work/R-2.8.1/src/main.
*** Error code 1

Stop in /usr/ports/math/R/work/R-2.8.1/src.
*** Error code 1

Stop in /usr/ports/math/R/work/R-2.8.1.
*** Error code 1

Stop in /usr/ports/math/R.
*** Error code 1

Stop in /usr/ports/math/R.
#

Seems like blas is to blame...

Here are some software version numbers:
blas-1.0_3
lapack-3.1.1_2
gcc-4.3.3_20090108
R-2.8.0

I haven't done any other other updates in the meantime (perl et. al.).
If I have the time I will check and see if my 8.0-testbox has the same 
problem (apart from being slow...).

Thanks for your time and help,
Philipp