[buildd] Known workaround for graphicsmagick failure?

[Daniel Kobras <kobras@debian.org>]

Hi!

graphicsmagick 1.1.7-6 failed to build on m68k due to what looks
suspiciously like a binutils bug:

cc -Wall -g -fno-strict-aliasing -O2 -Wall -o .libs/drawtest drawtest.o -pthread -L/usr/lib/X11 -L/usr/lib ../magick/.libs/libGraphicsMagick.so /usr/lib/liblcms.so /usr/lib/libtiff.so -lc /usr/lib/libfreetype.so /usr/lib/libjasper.so /usr/lib/libjpeg.so -lpng /usr/lib/libwmflite.so -lXext -lSM -lICE -lX11 -lbz2 /usr/lib/libxml2.so -lz -lm -lpthread -ldl
/usr/lib/libc_nonshared.a(elf-init.oS): In function `__libc_csu_fini': (.text+0x0): multiple definition of `__libc_csu_fini'
/usr/lib/libc_nonshared.a(elf-init.oS):(.text+0x0): first defined here
/usr/lib/libc_nonshared.a(elf-init.oS): In function `__libc_csu_init': (.text+0x52): multiple definition of `__libc_csu_init'
/usr/lib/libc_nonshared.a(elf-init.oS):(.text+0x52): first defined here
collect2: ld returned 1 exit status

The failing program 'drawtest' is just a part of the testsuite, so I
could just disable it on m68k an be done with it. However, I'd prefer to
keep regression tests in place. Is this a known bug already? Do you have
any suggestions on how to avoid it?

Thanks,

Daniel.