On Apr 11, 3:18 pm, Jason Moxham <[email protected]> wrote:
> On Saturday 11 April 2009 22:21:04 mabshoff wrote:
>
>
>
> > Opps,
>
> > this OSX 10.4 specific issues slipped through my testing so far. The
> > bug itself is generally present on OSX, but only leads to linking
> > failures on OSX 10.4.
>
> > g++ -dynamiclib ${wl}-undefined ${wl}dynamic_lookup -o .libs/
> > liblinboxsage.0.0.0.dylib .libs/linbox-sage.o -L/Users/Roed/Math/
> > sage-3.4.1.rc2/local/lib -lntl /Users/Roed/Math/sage-3.4.1.rc2/local/
> > lib/libgmpxx.dylib /Users/Roed/Math/sage-3.4.1.rc2/local/lib/
> > libmpir.dylib /Users/Roed/Math/sage-3.4.1.rc2/local/lib/libgmp.dylib /
> > Users/Roed/Math/sage-3.4.1.rc2/local/lib/libgivaro.dylib -L/Users/Roed/
> > Math/sage-3.4.1.rc2/local//lib -install_name /Users/Roed/Math/
> > sage-3.4.1.rc2/local/lib/liblinboxsage.0.dylib -compatibility_version
> > 1 -current_version 1.0 -Wl,-single_module
> > ld: multiple definitions of symbol ___gmpz_clear
> > /Users/Roed/Math/sage-3.4.1.rc2/local/lib/libmpir.dylib(clear.o)
> > definition of ___gmpz_clear
> > /Users/Roed/Math/sage-3.4.1.rc2/local/lib/libgmp.dylib(clear.o)
> > definition of ___gmpz_clear
> > <SNIP various other symbols>
> > ld: multiple definitions of symbol ___gmpz_sizeinbase
> > /Users/Roed/Math/sage-3.4.1.rc2/local/lib/libmpir.dylib(sizeinbase.o)
> > definition of ___gmpz_sizeinbase
> > /Users/Roed/Math/sage-3.4.1.rc2/local/lib/libgmp.dylib(sizeinbase.o)
> > definition of ___gmpz_sizeinbase
> > /usr/libexec/gcc/i686-apple-darwin8/4.0.1/libtool: internal link edit
> > command failed
> > make[4]: *** [liblinboxsage.la] Error 1
> > make[3]: *** [install-recursive] Error 1
> > make[2]: *** [install-recursive] Error 1
> > Error installing linbox
>
> > The problem lies in libgmpxx.la:
>
> > # Libraries that this one depends upon.
> > dependency_libs=' /tmp/mpr-1.0/lib/libmpir.la'
>
> > Because there is libmpir.la instead of libgmp.la linbox links with
>
> > -lgmpxx -mpir -lgmp
>
> > and things blow up as above. The manual workaround to make libgmpxx.la
> > depend on libgmp.la fixes the build.
>
> but then we get the problem , that it takes two go's to install it as
> before , as it wont install until it's dependences "are in".
> We could go back to sym-links , although would this have the same problem as
> above?
Hmm, I am not sure if symlinks would fix that. How about
"make install" installs mpir and all those bits
"make install-compat" installs all the *gmp* compat bits
install-compat would depend in install, so we would have a two stage
install.
Cheers,
Michael
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"mpir-devel" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/mpir-devel?hl=en
-~----------~----~----~----~------~----~------~--~---
