[Fink-users] Status on gcc43-4.3.0-1000?
There are at least two bugs in the linker from Xcode 3.1 which will prevent gcc43 from building. The first one manifests itself as a duplicate symbol in both tree-inline.o and libbackend.a (which contains tree-inline.o already) when linking cc1plus-dummy. This can be hacked around by removing the extra instance of tree-inline.o from gcc/cp/Make-lang.in. However this shouldn't be required since a unix linker should ignore duplicate symbols from a static library. Even with this workaround the build will fail later when linking parts of libjava. For now if you need gcc43 and the iPhoneSDK, I would revert to Xcode 3.0, build gcc43 and then reinstall the iPhoneSDK. Beware though that the Xcode 3.1 linker definitely has issues. Jack - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] php5 compile failure
Hi, I've recently had the need to recompile libapache2-mod-php5 version 5.1.4-122 but I am encountering an error during compilation. The error is that shown at http://paste.lisp.org/display/57503. It looks something like this: /bin/sh /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/libtool --silent --preserve-dup-deps --mode=compile gcc -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ -DPHP_ATOM_INC -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/include -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/main -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4 -I/sw/lib/system-openssl/include -I/sw/include/libxml2 -I/sw/include -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/date/lib -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/date/lib -I/usr/X11R6/include -I/usr/X11/include -I/usr/X11/include/freetype2 -I/sw/include/c-client -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/oniguruma -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/oniguruma -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/libmbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/libmbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/libmbfl/mbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/libmbfl/mbfl -I/sw/include/mysql -I/sw/include/postgresql -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/TSRM -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/Zend -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/main -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/TSRM -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ -DBIND_8_COMPAT -I/sw/include/apr-0 -I/sw/include/bind -I/sw/include -no-cpp-precomp -I/sw/include -L/sw/lib/system-openssl/lib -DZTS -c Zend/zend_language_parser.c -o Zend/zend_language_parser.lo i686-apple-darwin9-gcc-4.0.1: Zend/zend_language_parser.c: No such file or directory i686-apple-darwin9-gcc-4.0.1: no input files make: *** [Zend/zend_language_parser.lo] Error 1 ### execution of /var/tmp/tmp.1.fxNpgW failed, exit code 2 Removing runtime build-lock... Removing build-lock package... /sw/bin/dpkg-lockwait -r fink-buildlock-php5-5.1.4-122 (Reading database ... 26850 files and directories currently installed.) Removing fink-buildlock-php5-5.1.4-122 ... Failed: phase compiling: php5-5.1.4-122 failed I am on OS X 10.5 with fink version: [EMAIL PROTECTED]:~$ fink --version Package manager version: 0.28.1 Distribution version: selfupdate-rsync Wed Mar 19 12:31:39 2008, 10.5, i386 From the apache-build/Makefile generated during building, it looks like the line(s): Zend/zend_language_parser.lo: Zend/zend_language_parser.c $(LIBTOOL) --mode=compile $(CC) -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ $(COMMON_FLAGS) $(CFLAGS_CLEAN) $(EXTRA_CFLAGS) -c Zend/zend_language_parser.c -o Zend/zend_language_parser.lo Most other lines relating to stuff under Zend has a fully qualified path to the C source file, like this: Zend/zend_alloc.lo: /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/zend_alloc.c $(LIBTOOL) --mode=compile $(CC) -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ $(COMMON_FLAGS) $(CFLAGS_CLEAN) $(EXTRA_CFLAGS) -c /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/zend_alloc.c -o Zend/zend_alloc.lo I am not really all that familiar with the fink build system, so I haven't attempted to track down who is responsible for generating the faulty Makefile. Any input would be greatly appreciated. -- Søren O. Oh, bother said the Borg, we've assimilated Pooh. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] php5 compile failure
Søren Boll Overgaard wrote: Hi, I've recently had the need to recompile libapache2-mod-php5 version 5.1.4-122 but I am encountering an error during compilation. The error is that shown at http://paste.lisp.org/display/57503. It looks something like this: /bin/sh /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/libtool --silent --preserve-dup-deps --mode=compile gcc -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ -DPHP_ATOM_INC -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/include -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/main -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4 -I/sw/lib/system-openssl/include -I/sw/include/libxml2 -I/sw/include -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/date/lib -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/date/lib -I/usr/X11R6/include -I/usr/X11/include -I/usr/X11/include/freetype2 -I/sw/include/c-client -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/oniguruma -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/oniguruma -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/libmbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/libmbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/ext/mbstring/libmbfl/mbfl -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ext/mbstring/libmbfl/mbfl -I/sw/include/mysql -I/sw/include/postgresql -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/TSRM -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/Zend -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/main -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/TSRM -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/apache-build/ -DBIND_8_COMPAT -I/sw/include/apr-0 -I/sw/include/bind -I/sw/include -no-cpp-precomp -I/sw/include -L/sw/lib/system-openssl/lib -DZTS -c Zend/zend_language_parser.c -o Zend/zend_language_parser.lo i686-apple-darwin9-gcc-4.0.1: Zend/zend_language_parser.c: No such file or directory i686-apple-darwin9-gcc-4.0.1: no input files make: *** [Zend/zend_language_parser.lo] Error 1 ### execution of /var/tmp/tmp.1.fxNpgW failed, exit code 2 Removing runtime build-lock... Removing build-lock package... /sw/bin/dpkg-lockwait -r fink-buildlock-php5-5.1.4-122 (Reading database ... 26850 files and directories currently installed.) Removing fink-buildlock-php5-5.1.4-122 ... Failed: phase compiling: php5-5.1.4-122 failed I am on OS X 10.5 with fink version: [EMAIL PROTECTED]:~$ fink --version Package manager version: 0.28.1 Distribution version: selfupdate-rsync Wed Mar 19 12:31:39 2008, 10.5, i386 From the apache-build/Makefile generated during building, it looks like the line(s): Zend/zend_language_parser.lo: Zend/zend_language_parser.c $(LIBTOOL) --mode=compile $(CC) -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ $(COMMON_FLAGS) $(CFLAGS_CLEAN) $(EXTRA_CFLAGS) -c Zend/zend_language_parser.c -o Zend/zend_language_parser.lo Most other lines relating to stuff under Zend has a fully qualified path to the C source file, like this: Zend/zend_alloc.lo: /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/zend_alloc.c $(LIBTOOL) --mode=compile $(CC) -IZend/ -I/sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/ $(COMMON_FLAGS) $(CFLAGS_CLEAN) $(EXTRA_CFLAGS) -c /sw/src/fink.build/php5-5.1.4-122/php-5.1.4/Zend/zend_alloc.c -o Zend/zend_alloc.lo I am not really all that familiar with the fink build system, so I haven't attempted to track down who is responsible for generating the faulty Makefile. Any input would be greatly appreciated. I get the same problem on 10.5.2/PowerPC. I believe the Makefile is generated via options fed to ./configure, as usual--fink uses whatever mechanism the package itself provides. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Status on gcc43-4.3.0-1000?
On Wed, 19 Mar 2008 10:52:30 -0400 Jack Howarth [EMAIL PROTECTED] wrote: There are at least two bugs in the linker from Xcode 3.1 which will prevent gcc43 from building. The first one manifests itself as a duplicate symbol in both tree-inline.o and libbackend.a (which contains tree-inline.o already) when linking cc1plus-dummy. This can be hacked around by removing the extra instance of tree-inline.o from gcc/cp/Make-lang.in. However this shouldn't be required since a unix linker should ignore duplicate symbols from a static library. Even with this workaround the build will fail later when linking parts of libjava. For now if you need gcc43 and the iPhoneSDK, I would revert to Xcode 3.0, build gcc43 and then reinstall the iPhoneSDK. Beware though that the Xcode 3.1 linker definitely has issues. Jack Since gfortran is so important, and so slow and painful to compile, would it be possible for fink to host (or point to my) debian binaries for this? I do this myself, informally, for the benefit of my colleagues, and would be happy to make this available to anyone who is in need. Bill - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
In article [EMAIL PROTECTED], David Reiser [EMAIL PROTECTED] wrote: On Mar 16, 2008, at 9:38 PM, Richard Nelson wrote: I installed Gnucash2 and everything is working now except I cannot update stock prices. I ran /sw/bin/gnc-fq-update, and it reports no packages to install. I then ran /sw/bin/gnc-fq-dump, and it reports that Crypt::SSLeay is not installed and that I should run gnc-fq-update as root to install it. I rebuilt crypt-ssleay-pm588, no help. which perl gives me /opt/local/bin/perl. Is there something else I can do to get Finance::Quote working? Thanks! Richard That looks like macports' perl. Try hiding it (maybe by moving all of / opt/local to somewhere not in your path). I've patched the gnucash upstream gnc-fq-update so that it points at fink instead of CPAN, so when you try that, you certainly won't be installing anything that / opt/local/bin/perl can see. The other option is to find a way to get your default perl updated with Crypt::SSLeay. My experience has been that you can play games with perl behind gnucash's back, as long as every perl module you need is installed for whichever perl your system launches by default. Dave -- David Reiser [EMAIL PROTECTED] I thought I had removed Macports already due to a previous problem, but I did it again and now /opt/bin is gone. which perl now shows /usr/bin/perl. Running gnc-fq-dump still gives the alert about needing to install Crypt::SSLeay, but gnc-fq-update reports there is nothing to install. I installed Crypt::SSLeay using CPAN, rebuilt Gnucash, restarted, and still Finance::Quote reports that it is not loading properly. It is still probably some stupid mistake I have made, but I've hit the end of my troubleshooting skills. Richard - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Status on gcc43-4.3.0-1000?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 [EMAIL PROTECTED] wrote: | On Wed, 19 Mar 2008 10:52:30 -0400 | Jack Howarth [EMAIL PROTECTED] wrote: | | There are at least two bugs in the linker from Xcode 3.1 | which will prevent gcc43 from building. The first one manifests | itself as a duplicate symbol in both tree-inline.o and libbackend.a | (which contains tree-inline.o already) when linking cc1plus-dummy. | This can be hacked around by removing the extra instance of | tree-inline.o from gcc/cp/Make-lang.in. However this shouldn't | be required since a unix linker should ignore duplicate symbols | from a static library. Even with this workaround the build will | fail later when linking parts of libjava. For now if you need gcc43 | and the iPhoneSDK, I would revert to Xcode 3.0, build gcc43 | and then reinstall the iPhoneSDK. Beware though that the Xcode | 3.1 linker definitely has issues. | Jack | | | | Since gfortran is so important, and so slow and painful to compile, would it be possible for fink to host (or point to my) debian binaries for this? I do this myself, informally, for the benefit of my colleagues, and would be happy to make this available to anyone who is in need. | | Bill | | I've been using them myself. I saved myself several days of compiling :-) There has been discussion on the Fink IRC channel--it hasn't reached the point of actually needing to have a real paper trail yet--concerning ways to allow contributed binary distributions as official project elements. Official status would carry the implication that whomever was hosting such a repository can be trusted not to install arbitrary junk and thereby distribute broken binaries. :-) For now, we definitely can point people to your repository in an unofficial capacity, If you're willing and can stand the bandwidth. -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.7 (MingW32) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFH4XdRB8UpO3rKjQ8RAlVRAJ4lsFMnpEhLW/obTNarAvAvh4QcywCfT2/R El9Ka8CGkVSiOSYEYj7+YWk= =B5z6 -END PGP SIGNATURE- - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] php5 compile failure
Alexander Hansen wrote: I am not really all that familiar with the fink build system, so I haven't attempted to track down who is responsible for generating the faulty Makefile. Any input would be greatly appreciated. I get the same problem on 10.5.2/PowerPC. I believe the Makefile is generated via options fed to ./configure, as usual--fink uses whatever mechanism the package itself provides. Yup. configure gets most of its input from configure.in, and Makefile.am is used to generate the actual Makefile. But I have no idea what unpacks the source archive and ensures that ./configure gets called. Also my knowledge of autotools is just too basic to really get how Zend/Makefile.am is supposed to work. -- Søren O. Oh, bother said the Borg, we've assimilated Pooh. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] building fftw3 using gcc4.3 failed because of compatibility_version
Hi, For a few days now, the version 3.1.2-1 of fftw3 has been available. Building the package requires gcc4.3. The compilation ends up with the following error gcc -std=gnu99 -dynamiclib ${wl}-undefined ${wl}dynamic_lookup -o .libs/libfftw3_threads.3.1.2.dylib .libs/api.o .libs/conf.o .libs/threads.o .libs/dft-vrank-geq1.o .libs/ct.o .libs/rdft-vrank-geq1.o .libs/hc2hc.o .libs/vrank-geq1-rdft2.o .libs/f77api.o -L/sw/lib -lm -malign-double -install_name /sw/lib/libfftw3_threads.3.dylib -Wl,-compatibility_version -Wl,5 -Wl,-current_version -Wl,5.2 /usr/bin/ld: unknown flag: -compatibility_version I have to build the package manually and it turrns out that to solve this problem, it is sufficient to replace the file ltmain.sh from the fftw3 source by the one given by the package libtool14 (installed in /sw/share/libtool). Without this modification the generated version of libtool is not compatible with gcc4.3 concerning the compatibility_version flag. Actually the same problem occurs with libksba8. What is the good way to solve this issue ? Thanks for any help. Jérôme. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] building fftw3 using gcc4.3 failed because of compatibility_version
Jérôme LELONG wrote: Hi, For a few days now, the version 3.1.2-1 of fftw3 has been available. Building the package requires gcc4.3. The compilation ends up with the following error gcc -std=gnu99 -dynamiclib ${wl}-undefined ${wl}dynamic_lookup -o .libs/libfftw3_threads.3.1.2.dylib .libs/api.o .libs/conf.o .libs/threads.o .libs/dft-vrank-geq1.o .libs/ct.o .libs/rdft-vrank-geq1.o .libs/hc2hc.o .libs/vrank-geq1-rdft2.o .libs/f77api.o -L/sw/lib -lm -malign-double -install_name /sw/lib/libfftw3_threads.3.dylib -Wl,-compatibility_version -Wl,5 -Wl,-current_version -Wl,5.2 /usr/bin/ld: unknown flag: -compatibility_version I have to build the package manually and it turrns out that to solve this problem, it is sufficient to replace the file ltmain.sh from the fftw3 source by the one given by the package libtool14 (installed in /sw/share/libtool). Without this modification the generated version of libtool is not compatible with gcc4.3 concerning the compatibility_version flag. Actually the same problem occurs with libksba8. What is the good way to solve this issue ? Thanks for any help. Jérôme. libksba8 built fine for me, as did fftw3 (10.5.2) What OS version are you on, anyway? gcc43 in Fink doesn't install a gcc executable in the PATH to avoid colliding with that from Xcode. In any case, gcc from Xcode and that from gcc43 both appear to understand the -compatibility_version flag. Since the compiler line is just calling gcc let's make sure that it's doing the right thing. Can you provide the output from which gcc and gcc --version ? - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] building fftw3 using gcc4.3 failed because of compatibility_version
On 19 Mar 2008, at 22:03, Alexander Hansen wrote: Jérôme LELONG wrote: Hi, For a few days now, the version 3.1.2-1 of fftw3 has been available. Building the package requires gcc4.3. The compilation ends up with the following error gcc -std=gnu99 -dynamiclib ${wl}-undefined ${wl}dynamic_lookup -o .libs/libfftw3_threads.3.1.2.dylib .libs/api.o .libs/conf.o .libs/threads.o .libs/dft-vrank-geq1.o .libs/ct.o .libs/rdft-vrank-geq1.o .libs/hc2hc.o .libs/vrank-geq1-rdft2.o .libs/f77api.o -L/sw/lib -lm -malign-double -install_name /sw/lib/libfftw3_threads.3.dylib -Wl,-compatibility_version -Wl,5 -Wl,-current_version -Wl,5.2 /usr/bin/ld: unknown flag: -compatibility_version I have to build the package manually and it turrns out that to solve this problem, it is sufficient to replace the file ltmain.sh from the fftw3 source by the one given by the package libtool14 (installed in /sw/share/libtool). Without this modification the generated version of libtool is not compatible with gcc4.3 concerning the compatibility_version flag. Actually the same problem occurs with libksba8. What is the good way to solve this issue ? Thanks for any help. Jérôme. libksba8 built fine for me, as did fftw3 (10.5.2) What OS version are you on, anyway? gcc43 in Fink doesn't install a gcc executable in the PATH to avoid colliding with that from Xcode. In any case, gcc from Xcode and that from gcc43 both appear to understand the -compatibility_version flag. Right, but for ld the corresponding flag is -dylib_ compatibility_version... So this version of libtool is making an error prefixing - compatibility_version with _Wl, , which means pass directly to the linker For a way to fix this without a wholesale replacement of ltmain.sh, cf e.g. plplot.info, the lines # removing '${wl}' seems necessary (at least on 10.4) for gfortran linking, and to do no harm for the others .. sed -i.bak -e '/compatibility_/s,\${wl},,g' */ltmain.sh in the patchscript. JF Mertens - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
In article [EMAIL PROTECTED], Richard Nelson [EMAIL PROTECTED] wrote: In article [EMAIL PROTECTED], David Reiser [EMAIL PROTECTED] wrote: On Mar 16, 2008, at 9:38 PM, Richard Nelson wrote: I installed Gnucash2 and everything is working now except I cannot update stock prices. I ran /sw/bin/gnc-fq-update, and it reports no packages to install. I then ran /sw/bin/gnc-fq-dump, and it reports that Crypt::SSLeay is not installed and that I should run gnc-fq-update as root to install it. I rebuilt crypt-ssleay-pm588, no help. which perl gives me /opt/local/bin/perl. Is there something else I can do to get Finance::Quote working? Thanks! Richard That looks like macports' perl. Try hiding it (maybe by moving all of / opt/local to somewhere not in your path). I've patched the gnucash upstream gnc-fq-update so that it points at fink instead of CPAN, so when you try that, you certainly won't be installing anything that / opt/local/bin/perl can see. The other option is to find a way to get your default perl updated with Crypt::SSLeay. My experience has been that you can play games with perl behind gnucash's back, as long as every perl module you need is installed for whichever perl your system launches by default. Dave -- David Reiser [EMAIL PROTECTED] I thought I had removed Macports already due to a previous problem, but I did it again and now /opt/bin is gone. which perl now shows /usr/bin/perl. Running gnc-fq-dump still gives the alert about needing to install Crypt::SSLeay, but gnc-fq-update reports there is nothing to install. I installed Crypt::SSLeay using CPAN, rebuilt Gnucash, restarted, and still Finance::Quote reports that it is not loading properly. It is still probably some stupid mistake I have made, but I've hit the end of my troubleshooting skills. Richard I am finally able to get quotes. But so far the only way to do it is to start gnucash from the terminal with x11 already running. Starting from an x11 terminal means that Finance::Quote does not load properly. If I do /sw/bin/gnucash in terminal without x11 running it won't start at all. Richard - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Switch to gcc43 (Was Re: octave.info gcc-4.3)
On Mar 14, 2008, at 2:46 AM, Martin Costabel wrote: Jean-François Mertens wrote: [] but right, if true it would encourage a fast update of anything in fink to gcc43 I too have come round to the conclusion that this is indeed the best plan for action at this moment. I would suggest that we band together and do this update rather quickly. I am CCing fink-devel so that the maintainers of the 28 packages concerned get a chance to act, but I think we should just test this upgrade for each of these packages and then simply go ahead and change any package that builds correctly with gcc43. As far as I know, no problem with the switch from gcc42 to gcc43 has yet been reported. Here are the package descriptions concerned: graphics/pdftk.info libs/cfitsio.info sci/atlas.info sci/avl.info sci/cernlib2005.info sci/dx.info sci/fftw.info sci/fftw3.info sci/geant4.8.info sci/geant4.9.info sci/gopenmol.info sci/hdf5.info sci/libnc-dap3.info sci/netcdf-gfortran.info sci/octave-forge.info sci/octave.info sci/octplot.info sci/patchy4.info sci/plplot.info sci/qprop.info sci/root-pythia.info sci/root5-devel.info sci/root5.info sci/scilab.info sci/scipy-py.info sci/wip.info sci/xfoil.info utils/clamav.info -- Martin I continue to have problems with building the bleeding-edge octave sources. Unfortunately the bus error I encounter leaves gdb useless to me :-( In a nut shell, I'm unable to execute A\B or mldivide(A,B) in octave for more than 96 eqns/unknowns. It is has been pointed out that using gcc 4.0.1 (Apple's vecLib) will conflict with gfortran 4.1+, due to type mismatches resulting from how gfortran treats functions. I haven't looked at the detail, but when I googled, gfortran ff2c type mismatch quite a bit showed up. The suggested solution is to add -ff2c to the fortran compile parameters. I can't vouch for any of this, but if someone can tell me what to change in the atlas.info to pass -ff2c to gfortran, I can test suggested solution and report back. Ben - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] building fftw3 using gcc4.3 failed because of compatibility_version
Jean-François Mertens wrote: On 19 Mar 2008, at 22:03, Alexander Hansen wrote: Jérôme LELONG wrote: [] gcc -std=gnu99 -dynamiclib ${wl}-undefined ${wl}dynamic_lookup -o .libs/libfftw3_threads.3.1.2.dylib .libs/api.o .libs/conf.o .libs/threads.o .libs/dft-vrank-geq1.o .libs/ct.o .libs/rdft-vrank-geq1.o .libs/hc2hc.o .libs/vrank-geq1-rdft2.o .libs/f77api.o -L/sw/lib -lm -malign-double -install_name /sw/lib/libfftw3_threads.3.dylib -Wl,-compatibility_version -Wl,5 -Wl,-current_version -Wl,5.2 /usr/bin/ld: unknown flag: -compatibility_version I have to build the package manually and it turrns out that to solve this problem, it is sufficient to replace the file ltmain.sh from the fftw3 source by the one given by the package libtool14 (installed in /sw/share/libtool). Without this modification the generated version of libtool is not compatible with gcc4.3 concerning the compatibility_version flag. Actually the same problem occurs with libksba8. What is the good way to solve this issue ? Thanks for any help. Jérôme. libksba8 built fine for me, as did fftw3 (10.5.2) What OS version are you on, anyway? gcc43 in Fink doesn't install a gcc executable in the PATH to avoid colliding with that from Xcode. In any case, gcc from Xcode and that from gcc43 both appear to understand the -compatibility_version flag. Right, but for ld the corresponding flag is -dylib_ compatibility_version... So this version of libtool is making an error prefixing - compatibility_version with _Wl, , which means pass directly to the linker I find all this still very strange. I have been testing it on a small test case on 10.4.11 and am unable to produce this error. When I run gcc -dynamiclib, then the linker is not invoked as /usr/bin/ld directly, but as /usr/bin/libtool, and there the right name of the flag is indeed -compatibility_version (which on 10.5 is also understood by /usr/bin/ld). I have now also compiled fftw3 on 10.4.11, and I do not get the above error, although the compiler command line is exactly the same as for Jérôme. There must be something else going on. On what system versions are you getting these errors? And what does ld -v report there? -- Martin - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] building fftw3 using gcc4.3 failed because of compatibility_version
On 20 Mar 2008, at 01:04, Martin Costabel wrote: Jean-François Mertens wrote: On 19 Mar 2008, at 22:03, Alexander Hansen wrote: Jérôme LELONG wrote: [] /usr/bin/ld: unknown flag: -compatibility_version I have to build the package manually and it turrns out that to solve this problem, it is sufficient to replace the file ltmain.sh from the fftw3 source by the one given by the package libtool14 (installed in /sw/share/libtool). Without this modification the generated version of libtool is not compatible with gcc4.3 concerning the compatibility_version flag. Actually the same problem occurs with libksba8. What is the good way to solve this issue ? Thanks for any help. Jérôme. libksba8 built fine for me, as did fftw3 (10.5.2) What OS version are you on, anyway? gcc43 in Fink doesn't install a gcc executable in the PATH to avoid colliding with that from Xcode. In any case, gcc from Xcode and that from gcc43 both appear to understand the -compatibility_version flag. Right, but for ld the corresponding flag is -dylib_ compatibility_version... So this version of libtool is making an error prefixing - compatibility_version with _Wl, , which means pass directly to the linker I find all this still very strange. I have been testing it on a small test case on 10.4.11 and am unable to produce this error. When I run gcc -dynamiclib, then the linker is not invoked as / usr/bin/ld directly, but as /usr/bin/libtool, and there the right name of the flag is indeed -compatibility_version (which on 10.5 is also understood by /usr/bin/ld). this is correct; for me too (on 10.4). That is why the fix in plplot was preceded by the comment: # removing '${wl}' seems necessary (at least on 10.4) for gfortran linking, and to do no harm for the others .. stressing gfortran, which didn't call libtool, but either collect2 or ld directly, if I recall correctly. (running the command with '-v') I have now also compiled fftw3 on 10.4.11, and I do not get the above error, although the compiler command line is exactly the same as for Jérôme. Same experience here.. There must be something else going on. Different compilers, different linkers.. (lost a bit track of trying to understand what goes on in 10.5, especially with all this iPhone (?) business recently..) On what system versions are you getting these errors? And what does ld -v report there? cctools-622.8~4 JF - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
On Mar 19, 2008, at 7:34 PM, Richard Nelson wrote: snip I am finally able to get quotes. But so far the only way to do it is to start gnucash from the terminal with x11 already running. Starting from an x11 terminal means that Finance::Quote does not load properly. If I do /sw/bin/gnucash in terminal without x11 running it won't start at all. Richard This may be the standard issue with X11.app not processing the user environment customizations. I have my . /sw/bin/init.sh in ~/.bashrc instead of ~/.profile or ~/.bash_profile because that's one of the startup files that Apple's X11 knows about. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
On Mar 19, 2008, at 8:50 PM, Alexander Hansen wrote: On Mar 19, 2008, at 7:34 PM, Richard Nelson wrote: snip I am finally able to get quotes. But so far the only way to do it is to start gnucash from the terminal with x11 already running. Starting from an x11 terminal means that Finance::Quote does not load properly. If I do /sw/bin/gnucash in terminal without x11 running it won't start at all. Richard This may be the standard issue with X11.app not processing the user environment customizations. I have my . /sw/bin/init.sh in ~/.bashrc instead of ~/.profile or ~/.bash_profile because that's one of the startup files that Apple's X11 knows about. mine's in ~/.profile and not in ~/.bashrc, so while it might help, it doesn't appear to be required. What happens if you just type 'gnucash' in a terminal window? If you have to use /sw/bin/gnucash, then I think you should try running /sw/bin/pathsetup.sh to see if that helps. The gnc-fq-helper that launches finance-quote for gnucash is in /sw/bin, so your system probably looks in the current app's directory as part of the ephemeral path once x11 knows where gnucash came from. You should only have to run pathsetup once. If that doesn't work, please post the output of: fink --version | head -n 2 and fink list -i gnucash2 finance-quote libwww html-tableextract crypt- ssleay perl (that last bit is one line, but the mailing list may wrap it). Also, do you get the same answer to 'which perl' when you type it in a regular terminal and in an xterm window? (I sure hope so, but I'm fishin' here...) Dave -- David Reiser [EMAIL PROTECTED] - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Switch to gcc43 (Was Re: octave.info gcc-4.3)
Ben Abbott wrote: On Mar 14, 2008, at 2:46 AM, Martin Costabel wrote: Jean-François Mertens wrote: [] but right, if true it would encourage a fast update of anything in fink to gcc43 I too have come round to the conclusion that this is indeed the best plan for action at this moment. I would suggest that we band together and do this update rather quickly. I am CCing fink-devel so that the maintainers of the 28 packages concerned get a chance to act, but I think we should just test this upgrade for each of these packages and then simply go ahead and change any package that builds correctly with gcc43. As far as I know, no problem with the switch from gcc42 to gcc43 has yet been reported. Here are the package descriptions concerned: graphics/pdftk.info libs/cfitsio.info sci/atlas.info sci/avl.info sci/cernlib2005.info sci/dx.info sci/fftw.info sci/fftw3.info sci/geant4.8.info sci/geant4.9.info sci/gopenmol.info sci/hdf5.info sci/libnc-dap3.info sci/netcdf-gfortran.info sci/octave-forge.info sci/octave.info sci/octplot.info sci/patchy4.info sci/plplot.info sci/qprop.info sci/root-pythia.info sci/root5-devel.info sci/root5.info sci/scilab.info sci/scipy-py.info sci/wip.info sci/xfoil.info utils/clamav.info -- Martin I continue to have problems with building the bleeding-edge octave sources. Unfortunately the bus error I encounter leaves gdb useless to me :-( In a nut shell, I'm unable to execute A\B or mldivide(A,B) in octave for more than 96 eqns/unknowns. It is has been pointed out that using gcc 4.0.1 (Apple's vecLib) will conflict with gfortran 4.1+, due to type mismatches resulting from how gfortran treats functions. I haven't looked at the detail, but when I googled, gfortran ff2c type mismatch quite a bit showed up. The suggested solution is to add -ff2c to the fortran compile parameters. I can't vouch for any of this, but if someone can tell me what to change in the atlas.info to pass -ff2c to gfortran, I can test suggested solution and report back. Ben Update, I built octave using gfortran with -ff2c added to the configure parameters and passed all tests included in make check ./configure '--prefix=/sw' 'FLIBS=/sw/lib/gcc4.3/lib/libgfortran.dylib' 'F77=/sw/bin/gfortran' '--libexecdir=${prefix}/lib' 'CFLAGS=-g -O3' 'LDFLAGS=-g -L/sw/lib' 'CPPFLAGS=-I/sw/include' 'CXXFLAGS=-g -O3' 'FFLAGS=-g -O3 -fbounds-check -ff2c' '--without-blas' '--without-lapack' The lapack and blas were satisfied by the versions included in Octave's libcruft. Ben -- View this message in context: http://www.nabble.com/octave.info---gcc-4.3-tp16027509p16169677.html Sent from the fink-users mailing list archive at Nabble.com. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
[Fink-users] The process has forked
Gentemen, Some time ago (just after Leopard came out) the below discussion took place on this list regarding CoreFoundation and how changes in Leopard impact some application. I've been able to google the discussion, but I haven't seen if there is any resolution. I have the same symptoms, particularly when running startkde I'm able to get kde up and running more or less by running /sw/bin/kicker, but a lot doesn't function -- such as sound -- the way it did under Tiger. Any ideas? Or am I just pretty much out of luck for now? Best regards, Charles -- Charles Sheehan-Miles http://www.sheehanmiles.com | email: [EMAIL PROTECTED] - -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Richard E. Miles wrote: snip __THE_PROCESS_HAS_FORKED_AND_YOU_CANNOT_USE_THIS_COREFOUNDATION_FUNCTIONALITY___ YOU_MUST_EXEC__ () to debug. The process has forked and you cannot use this CoreFoundation functionality safely. You MUST exec(). Break on __THE_PROCESS_HAS_FORKED_AND_YOU_CANNOT_USE_THIS_COREFOUNDATION_FUNCTI snip That's a known problem, and the maintainer is working on it. It appears that apps other than the desktop don't suffer from this, so you might want to get by with not using startkde for now and just using the individual applications as needed. Okay thanks for the answer. Do you think that we will have a solution quickly or will it be down the line and later? Probably not very quickly. The maintainer has mentioned having severe time constraints for doing Fink stuff, and he is right now involved with getting the online package database back up and running. After that there is a major GNOME update, and then he will have time for his own packages. - -- -- - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
In article [EMAIL PROTECTED], David Reiser [EMAIL PROTECTED] wrote: On Mar 19, 2008, at 8:50 PM, Alexander Hansen wrote: On Mar 19, 2008, at 7:34 PM, Richard Nelson wrote: snip I am finally able to get quotes. But so far the only way to do it is to start gnucash from the terminal with x11 already running. Starting from an x11 terminal means that Finance::Quote does not load properly. If I do /sw/bin/gnucash in terminal without x11 running it won't start at all. Richard This may be the standard issue with X11.app not processing the user environment customizations. I have my . /sw/bin/init.sh in ~/.bashrc instead of ~/.profile or ~/.bash_profile because that's one of the startup files that Apple's X11 knows about. mine's in ~/.profile and not in ~/.bashrc, so while it might help, it doesn't appear to be required. What happens if you just type 'gnucash' in a terminal window? If you have to use /sw/bin/gnucash, then I think you should try running /sw/bin/pathsetup.sh to see if that helps. The gnc-fq-helper that launches finance-quote for gnucash is in /sw/bin, so your system probably looks in the current app's directory as part of the ephemeral path once x11 knows where gnucash came from. You should only have to run pathsetup once. If that doesn't work, please post the output of: fink --version | head -n 2 and fink list -i gnucash2 finance-quote libwww html-tableextract crypt- ssleay perl (that last bit is one line, but the mailing list may wrap it). Also, do you get the same answer to 'which perl' when you type it in a regular terminal and in an xterm window? (I sure hope so, but I'm fishin' here...) Dave -- David Reiser [EMAIL PROTECTED] I can type either gnucash or sw/bin/gnucash in a terminal window and gnucash will start with Finance::Quote properly loaded, but only if x11 is already running. If x11 is not running it will not start at all when I type either of these commands into terminal. Which perl produces usr/bin/perl in either a terminal window or an x11 terminal window. I am running package manager 0.28.1. Here is the result of fink list -i gnucash2 finance-quote libwww html-tableextract crypt-ssleay perl: i crypt-ssleay-pm 0.51-2 Perl module, allows perl to use OpenSSL i finance-quote-p 1.13-3 Perl module to get stock quotes i finance-quotehi 1.11-1 Perl module to get old financial quotes i gnucash2 2.2.4-101Financial-accounting software i gnucash2-shlibs 2.2.4-101Financial-accounting software i html-tableextra 2.10-1 Perl module to extract data from html tables i libwww 5.4.0-7 General-purpose cross-platform Web API in C i libwww-bin 5.4.0-7 General-purpose cross-platform Web API in C i libwww-pm588 5.805-2 Perl modules, programming interface to WWW i libwww-shlibs5.4.0-7 General-purpose cross-platform Web API in C p perl588-core [virtual package] p perlio-eol-pm58 [virtual package] p perlobjcbridge- [virtual package] i system-perl 5.8.8-1 [virtual package representing perl] p system-perl588[virtual package] Thanks for all the help. I'm pretty happy with the way things are working now that I'm back to a working Finance::Quote install. It would be nice to get things working perfectly though. Richard - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] Switch to gcc43 (Was Re: octave.info gcc-4.3)
Ben Abbott wrote: Ben Abbott wrote: On Mar 14, 2008, at 2:46 AM, Martin Costabel wrote: Jean-François Mertens wrote: [] but right, if true it would encourage a fast update of anything in fink to gcc43 I too have come round to the conclusion that this is indeed the best plan for action at this moment. I would suggest that we band together and do this update rather quickly. I am CCing fink-devel so that the maintainers of the 28 packages concerned get a chance to act, but I think we should just test this upgrade for each of these packages and then simply go ahead and change any package that builds correctly with gcc43. As far as I know, no problem with the switch from gcc42 to gcc43 has yet been reported. Here are the package descriptions concerned: graphics/pdftk.info libs/cfitsio.info sci/atlas.info sci/avl.info sci/cernlib2005.info sci/dx.info sci/fftw.info sci/fftw3.info sci/geant4.8.info sci/geant4.9.info sci/gopenmol.info sci/hdf5.info sci/libnc-dap3.info sci/netcdf-gfortran.info sci/octave-forge.info sci/octave.info sci/octplot.info sci/patchy4.info sci/plplot.info sci/qprop.info sci/root-pythia.info sci/root5-devel.info sci/root5.info sci/scilab.info sci/scipy-py.info sci/wip.info sci/xfoil.info utils/clamav.info -- Martin I continue to have problems with building the bleeding-edge octave sources. Unfortunately the bus error I encounter leaves gdb useless to me :-( In a nut shell, I'm unable to execute A\B or mldivide(A,B) in octave for more than 96 eqns/unknowns. It is has been pointed out that using gcc 4.0.1 (Apple's vecLib) will conflict with gfortran 4.1+, due to type mismatches resulting from how gfortran treats functions. I haven't looked at the detail, but when I googled, gfortran ff2c type mismatch quite a bit showed up. The suggested solution is to add -ff2c to the fortran compile parameters. I can't vouch for any of this, but if someone can tell me what to change in the atlas.info to pass -ff2c to gfortran, I can test suggested solution and report back. Ben Update, I built octave using gfortran with -ff2c added to the configure parameters and passed all tests included in make check ./configure '--prefix=/sw' 'FLIBS=/sw/lib/gcc4.3/lib/libgfortran.dylib' 'F77=/sw/bin/gfortran' '--libexecdir=${prefix}/lib' 'CFLAGS=-g -O3' 'LDFLAGS=-g -L/sw/lib' 'CPPFLAGS=-I/sw/include' 'CXXFLAGS=-g -O3' 'FFLAGS=-g -O3 -fbounds-check -ff2c' '--without-blas' '--without-lapack' The lapack and blas were satisfied by the versions included in Octave's libcruft. Ben I've just completed a build using the configure options passed by octave.info, except that -ff2c was included in the FFLAGS. make check passed all tests! Ben -- View this message in context: http://www.nabble.com/octave.info---gcc-4.3-tp16027509p16170713.html Sent from the fink-users mailing list archive at Nabble.com. - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users
Re: [Fink-users] gnucash can't get stock prices
On Mar 19, 2008, at 10:43 PM, Richard Nelson wrote: I can type either gnucash or sw/bin/gnucash in a terminal window and gnucash will start with Finance::Quote properly loaded, but only if x11 is already running. If x11 is not running it will not start at all when I type either of these commands into terminal. which version of 10.5, and what does the About X11 dialog say for xquartz and xorg server versions? Dave -- David Reiser [EMAIL PROTECTED] - This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/ ___ Fink-users mailing list Fink-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/fink-users