On Wednesday 02 February 2005 21:22, Simon Edwards wrote: > It appears that libxf86config.a can't be used by shared libraries. It > doesn't appear to have been compiled correctly using the -fPIC compile > option. > > I'm currently developing a util in Python that uses has a C extension which > uses libxf86config.a. In terms of libraries and linking, we are building a > small *.so shared library that links with libxf86config.a. Linking is > failing: > > ------------------------------------- > building 'ixf86config' extension > gcc -pthread -shared build/temp.linux-x86_64-2.3/modules/xf86config_ext.o > -L/usr/X11R6/lib -lX11 -lXxf86vm -lXext -lglib -lxf86config -o > build/lib.linux-x86_64-2.3/ixf86config.so > /usr/bin/ld: /usr/X11R6/lib/libxf86config.a(scan.o): relocation R_X86_64_32 > can not be used when making a shared object; recompile with -fPIC > /usr/X11R6/lib/libxf86config.a: could not read symbols: Bad value > collect2: ld returned 1 exit status > error: command 'gcc' failed with exit status 1 > ------------------------------------- > > We've also tried linking against libxf86config_pic.a. The result was the > same. In fact libxf86config_pic.a and libxf86config.a were identical files. > > Looking into libxf86config.a using objdump, I see that scan.o does indeed > have a number of R_X86_64_32 relocations. What I also noticed is that the > other objects in libxf86config.a *don't* have any R_X86_64_32 relocations > at all. Did something go wrong when scan.o was compiled?
The packages involved are: xlibs 4.3.0.dfsg.1-10 xlibs-data 4.3.0.dfsg.1-10 xlibs-dev 4.3.0.dfsg.1-10 xlibs-pic 4.3.0.dfsg.1-10 xlibs-static-dev 4.3.0.dfsg.1-10 xlibs-static-pic 4.3.0.dfsg.1-10 They are all from the pure64 packages on a mirrorsite of debian-amd64.alioth.debian.org. I can confirm that the same builds fine with i386-sid. Please don't hesitate to contact me in case you need more details. To reproduce it, download http://www.simonzone.com/software/guidance/guidance-0.2.0.tar.bz2 and try to build ixf86config (see above for gcc parameters). Could someone have a look at it? Thanks, sebas -- http://vizZzion.org ====================== Everyone complains of his memory, no one of his judgement. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
