Package: kernel-package Version: 10.011 Severity: minor Installing a linux-image built on sid with kernel-package 10.011 fails on a sarge system using lilo. It looks like the postinst script doesn't properly handle lilo's zero return code. I have added debug output to the postinst:
if (-T "/etc/$loader.conf") { # Trust and use the existing lilo.conf. print "You already have a $Loader configuration in /etc/$loader.conf\n"; warn "use lilo.conf\n"; my $ret = &run_lilo(); warn "ret $ret\n"; exit $ret; warn "after exit\n"; } and here is the output of that script $ DEBUG=yes sudo /var/lib/dpkg/info/linux-image-2.6.14.3-zgsrv.postinst configure vmlinuz(/boot/vmlinuz-2.6.14.3-zgsrv) points to /boot/vmlinuz-2.6.14.3-zgsrv (/boot/vmlinuz-2.6.14.3-zgsrv) -- doing nothing at /var/lib/dpkg/info/linux-image-2.6.14.3-zgsrv.postinst line 580, <STDIN> line 2. use lilo.conf ret 0 $ echo $? 128 $ less This, of course, makes aptitude barf. What I find strange is, that an explicit "exit 0" right after the "run_lilo" call makes the shell say "128", and an explicit "exit 0" right before the "run_lilo" makes the shell say "0" as it should be. This is very strange. Any ideas? I inspected the changelog of current kernel_package and didn't find any entries that indicate that this issue was fixed in more recent versions of kernel-package. Greetings Marc -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]