>Seems that the glibc 2.2.2-4 that just went into unstable exhibits this "cpp0 >dumps core" problem again. Those of you running woody might want to pin your >libc6 at whatever the previous version was until we can work out what's wrong.
There seems to be some kind of rlimit handling fuck-up. The new glibc package was built against 2.4 kernel headers, and it works fine if you're actually running a 2.4 kernel but fails under 2.2.19. I think cpp is dying because it accidentally sets itself a ludicrously restrictive stack limit. More news later. p.
