[Rd] La_svd compiler error

Antczak, Philipp P.Antczak at liverpool.ac.uk
Fri Apr 12 13:21:51 CEST 2013


I did include the Makevars in the src directory with the following line: 
PKG_LIBS = $(LAPACK_LIBS) $(BLAS_LIBS) $(FLIBS)

But this had no effect at all - still failed at La_svd
As I said in Linux no problem at all - only in windows.

Thanks
Phil

-----Original Message-----
From: Duncan Murdoch [mailto:murdoch.duncan at gmail.com] 
Sent: 12 April 2013 12:13
To: Antczak, Philipp
Cc: 'r-devel at R-project.org'
Subject: Re: [Rd] La_svd compiler error

On 13-04-11 7:41 AM, Antczak, Philipp wrote:
> Dear All,
>
> I have been trying to compile a package for windows that we have written. Since R version 3.0.0 the package doesn't compile any more. (On Linux the package compiles without errors).
> The error specifically says:
>
> gcc -m32 -I"C:/PROGRA~1/R/R-30~1.0/include" -DNDEBUG     -I"d:/RCompile/CRANpkg/extralibs64/local/include"     -O3 -Wall  -std=gnu99 -mtune=core2 -c galgoDistance.c -o galgoDistance.o
> gcc -m32 -I"C:/PROGRA~1/R/R-30~1.0/include" -DNDEBUG     -I"d:/RCompile/CRANpkg/extralibs64/local/include"     -O3 -Wall  -std=gnu99 -mtune=core2 -c galgo_fitnesses.c -o galgo_fitnesses.o
> gcc -m32 -shared -s -static-libgcc -o galgo.dll tmp.def galgoDistance.o galgo_fitnesses.o -LC:/PROGRA~1/R/R-30~1.0/bin/i386 -lRlapack -LC:/PROGRA~1/R/R-30~1.0/bin/i386 -lRblas -lgfortran -Ld:/RCompile/CRANpkg/extralibs64/local/lib/i386 -Ld:/RCompile/CRANpkg/extralibs64/local/lib -LC:/PROGRA~1/R/R-30~1.0/bin/i386 -lR
> galgo_fitnesses.o:galgo_fitnesses.c:(.text+0xda1): undefined reference to `La_svd'
> collect2: ld returned 1 exit status
>
> specifically from the error I understand that error occurs in the link to La_svd which is defined in the first couple of lines in the code:
>
> #include <math.h>
> #include <R.h>
> #include <Rinternals.h>
> #include <Rmath.h>
>
> SEXP La_svd(SEXP jobu, SEXP jobv, SEXP x, SEXP s, SEXP u, SEXP v, SEXP method);
>

You reported this as a bug, and were told what to do to try to fix it. 
Did that work?

Duncan Murdoch



More information about the R-devel mailing list