On Wed, Oct 30 2002, Goto Masanori wrote: > I notice that you use "/usr/lib/libcrypto.so.0.9.6", but in my machine > it uses "/usr/lib/i686/libcrypto.so.0.9.6". libssl has architecture > optimization package, is it possibly related issue? Since Adam's machine is a powerpc, it would be a bad thing if his linker tried to load anything from /usr/lib/i686. ;) Adam tells me it should be easy for me to reproduce this bug in a clean sid chroot. Unfortunately, on my devel machine (where I can't reproduce the segfault), I can't *install* a clean sid chroot, due to a bus error when trying to install one of the packages. I can only conclude that this is not a bug; it's a curse. I've looked for common factors between the bug reports, that distinguish the submitters' configuration from my own. I have ruled out: individual uploads of glibc 2.3.1 (the problem is reported on 1, 2, and 3) kernel series (the problem is reported on 2.2 and 2.4; I'm running 2.4.16) SASL configuration (I've wiped my SASL configs and started from scratch) processor type (I have a Pentium II; problem is reported on both x86 AND ppc) I have also confirmed that I cannot reproduce the bug when all of my binary packages for php4-imap's exactly match the versions in one of the bug reports. It's definitely a curse. Steve Langasek postmodern programmer
Attachment:
pgpSeBJDT6TSp.pgp
Description: PGP signature