> Yes, libc6-i386 is the package that "owns" /usr/lib32, so I believe this is > a bug in lib32gfortran0. It looks like this bug can be fixed for etch > simply by letting the new version of gcc-4.0 in, which no longer builds > lib32gfortran0 at all; the lib32gfortran1 package from gcc-4.1 does not have > this bug. yes, letting gcc-4.0 into testing would be appreciated.