That would mean that either:
- one of our local patches is broken
- your binary on the disk is broken (check with rpm -Va the md5sums)
- you used a different compiler than we did - but we ship the same that we use...
I'm really surprised that recompilation helped...
... - or it was a placebo effect. Not uncommon in software testing.
I guess you could argue that crashing consistently and then not crashing after recompilation has some placebo in it ;) More likely it was version differences; I obtained the source from the project pages directly and did not consider checking whether the latest version was the same as the SUSE rpms. In some cases I had to play with configure options like --disable-mmx and so forth. Not very scientific, and I didn't make any notes ... just stuck with the policy of disabling all platform-specific options. All part of a pointless effort anyways, since I had to go back to a 32-bit linux kernel/distro just so I could run VMWare for testing kmods... _m _________________________________________________________________ Rethink your business approach for the new year with the helpful tips here. http://special.msn.com/bcentral/prep04.armx