With Alex's latest fixes, gsl-oct324 now builds for me against octave324-atlas on powerpc-darwin8. (It also built fine against octave324 non-atlas before.)
Did anyone happen to notice during octave324-atlas's build: http://paste.lisp.org/display/124633 configure:13367: checking for F77_FUNC(dseupd,DSEUPD) in -larpack configure:13400: flag-sort -r gcc -o conftest -g -O3 -MD -I/sw/include -I/sw/include -L/sw/lib conftest.c -larpack /sw/lib/liblapack.dylib /sw/lib/gcc4.6/lib/libgfortran.dylib -lhdf5 -lz -lm -lGraphicsMagick -lmetis >&5 /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: warning multiple definitions of symbol _second_ /sw/lib/liblapack.dylib(single module) definition of _second_ /sw/lib/libarpack.dylib(second.o) definition of _second_ /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: warning suggest use of -bind_at_load, as lazy binding may result in errors or different symbols being used /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: Undefined symbols: _dcopy_ _dger_ _dnrm2_ _dscal_ _dswap_ symbol _second_ used from dynamic library /sw/lib/liblapack.dylib(single module) not from earlier dynamic library /sw/lib/libarpack.0.dylib(second.o) Fang >> Hi, >> I've fixed octave-atlas so that CDOTU and ZDOTU are now called with >> the correct convention for the (fink-built) atlas variant. >> However, now when I rebuild gsl-oct324 (against octave324-atlas), I >> see your undefined linker errors. This is a separate issue. > > I see: > mkoctfile -DHAVE_OCTAVE_32 -v coupling_3j.cc -lgsl > /sw/var/lib/fink/path-prefix-g++-4.0/g++ -c -I/sw/include -I/sw/include > -I/sw/include/octave-3.2.4 -I/sw/include/octave-3.2.4/octave -I/sw/include > -I/sw/include/freetype2 -I/sw/include -g -O3 -MD -DHAVE_OCTAVE_32 > coupling_3j.cc -o coupling_3j.o > /sw/var/lib/fink/path-prefix-g++-4.0/g++ -bundle -bundle_loader > /sw/bin/octave-3.2.4 -o coupling_3j.oct coupling_3j.o -lgsl > -L/sw/lib/octave-3.2.4 -L/sw/lib -loctinterp -loctave -lcruft -L/sw/lib > /sw/lib/liblapack.dylib /sw/lib/libf77blas.dylib -lfftw3 -lfftw3f -lreadline > -lncurses -lhdf5 -lz -lm -lGraphicsMagick -lmetis > /sw/lib/gcc4.6/lib/libgfortran.dylib > > before the link-fail. > > These symbols: > _ cblas_caxpy > _ cblas_ccopy > ... > > are defined in %p/lib/libcblas.dylib, which doesn't appear on the > command-line linker flags. Doh! > I think akh is working on a fix as we read this. :) > > Fang > >> > > During my testing, I found that I *needed* -ff2c. >> > > I'll look into this again later today, now that I have atlas built. >> > > The -ff2c should be conditional on whether system-atlas or fink-atlas >> > > is >> > > used. >> > >> > This is what I guessed from the comments. >> > >> > > I'm away from computer the rest of today, but will get back to it as >> > > son >> > > as I can. >> > >> > The only emergency is that I report what I find before I forget it!-) >> > My life no longer depends on my G4. >> > >> > > in the meantime, just remove the '-ff2c' flags from the FFLAGS >> > > in the info file, >> > > and do the same if needed for arpack and qrupdate, then >> > > rebuild first the latter 2,then octve >> > >> > This what I have done. After that mkoctfile -p BLAS_LIBS returns >> > /sw/lib/liblapack.dylib /sw/lib/libf77blas.dylib >> > but install gsl-oct324 still fails with: >> > >> > ... >> > find . -name NOINSTALL -print # shows which toolboxes won't be >> > installed >> > >> > /usr/libexec/gcc/powerpc-apple-darwin8/4.0.1/ld: Undefined symbols: >> > _ cblas_caxpy >> > _ cblas_ccopy >> > ... >> > collect2: ld returned 1 exit status >> > make: *** [coupling_3j.oct] Error 1 >> > 'make' returned the following error: /bin/sh buildgsl_sf.sh >> > mkoctfile -DHAVE_OCTAVE_32 -v coupling_3j.cc -lgsl >> > /sw/var/lib/fink/path-prefix-g++-4.0/g++ -c -I/sw/include -I/sw/include >> > -I/sw/include/octave-3.2.4 -I/sw/include/octave-3.2.4/octave >> > -I/sw/include >> > -I/sw/include/freetype2 -I/sw/include -g -O3 -MD -DHAVE_OCTAVE_32 >> > coupling_3j.cc -o coupling_3j.o >> > /sw/var/lib/fink/path-prefix-g++-4.0/g++ -bundle -bundle_loader >> > /sw/bin/octave-3.2.4 -o coupling_3j.oct coupling_3j.o -lgsl >> > -L/sw/lib/octave-3.2.4 -L/sw/lib -loctinterp -loctave -lcruft -L/sw/lib >> > /sw/lib/liblapack.dylib /sw/lib/libf77blas.dylib -lfftw3 -lfftw3f >> > -lreadline -lncurses -lhdf5 -lz -lm -lGraphicsMagick -lmetis >> > /sw/lib/gcc4.6/lib/libgfortran.dylib >> > error: called from `pkg>configure_make' in file >> > error: /sw/share/octave/3.2.4/m/pkg/pkg.m near line 1253, column 2 >> > error: called from: >> > error: /sw/share/octave/3.2.4/m/pkg/pkg.m at line 714, column 5 >> > error: /sw/share/octave/3.2.4/m/pkg/pkg.m at line 287, column 7 >> > error: bld/octave-forge-comp at line 4, column 1 >> > ### execution of /sw/share/octave/3.2.4/scripts/octave-forge-compile.sh >> > failed, exit code 1 >> > Removing runtime build-lock... >> > Removing build-lock package... >> > /sw/bin/dpkg-lockwait -r fink-buildlock-gsl-oct324-1.0.8-1 >> > (Reading database ... 469438 files and directories currently >> > installed.) >> > Removing fink-buildlock-gsl-oct324-1.0.8-1 ... >> > Failed: phase compiling: gsl-oct324-1.0.8-1 failed >> > >> > So I decided to rebuild fltk-backend-aqua-oct324 and I saw >> > >> > ... >> > checking whether CDOTU is called correctly from Fortran... no >> > checking whether ZDOTU is called correctly from Fortran... no >> > configure: WARNING: A BLAS library was detected but found incompatible >> > with your Fortran 77 compiler. The reference BLAS implementation will >> > be >> > used. To improve performance, consider using a different Fortran >> > compiler >> > or a switch like -ff2c to make your Fortran compiler use a calling >> > convention compatible with the way your BLAS library was compiled, or >> > use >> > a different BLAS library. >> > ... >> > >> > and indeed gsl-oct324-1.0.8-1 failed again. The offending test is >> > >> > configure: 12141: /sw/bin/gfortran-fsf-4.6 -o conftest -O3 -L/sw/lib >> > configure: conftest.f >> > configure: >> > -Wl,-framework,Accelerate,-dylib_file,/System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib:/System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib >> > >> > configure: -lhdf5 -lz -lm -lGraphicsMagick -lmetis >&5 >> > configure: 12141: $? = 0 >> > configure: 12141: ./conftest >> > ./configure: line 1: 2554 Segmentation fault ./conftest$ac_exeext >> > >> > for CDOTU. It seems that atlas is not taken into account and that -ff2c >> > is >> > required without it. Not that I grepped >> > /sw/fink/10.4/unstable/main/finkinfo/10.4-EOL/sci/* >> > for ff2c and the only occurences reported were in comments. >> > >> > Cheers, >> > >> > Dominique >> > >> >> > > -- David Fang http://www.csl.cornell.edu/~fang/ http://www.achronix.com/ ------------------------------------------------------------------------------ BlackBerry® DevCon Americas, Oct. 18-20, San Francisco, CA Learn about the latest advances in developing for the BlackBerry® mobile platform with sessions, labs & more. See new tools and technologies. Register for BlackBerry® DevCon today! http://p.sf.net/sfu/rim-devcon-copy1 _______________________________________________ Fink-users mailing list Fink-users@lists.sourceforge.net List archive: http://news.gmane.org/gmane.os.macosx.fink.user Subscription management: https://lists.sourceforge.net/lists/listinfo/fink-users
