Bug#401331: initramfs-tools: lilo failures in update-initramfs are silent.

[maximilian attems <maks@sternwelten.at>]

On Sat, Dec 02, 2006 at 10:06:21AM -0700, David Anselmi wrote:
> 
> When lilo -t fails update-initramfs exits with an error (due to set -e)
> but provides no indication to the user.  When called in postinst the
> failure is reported by, e.g., aptitude but there is no explanation of
> what went wrong.
> 
> This could leave a system unbootable if update-initramfs is run outside
> of package management (where the failure is invisible) or if the user doesn't
> realize an upgrade failure might be serious.
> 
> I suggest that the stderr of lilo -t should at least be provided to the
> user, or perhaps run_lilo() should do set +e and then complain loudly if
> lilo -t fails.

yup the failure needs to be reported,
fully agreed - i'll refix that for 0.85d
 
> -- Package-specific info:
> -- /proc/cmdline
> auto BOOT_IMAGE=Linux ro root=fd00
> 
> -- /proc/filesystems
> 	cramfs
> 	ext3
> 
> -- lsmod
> [ Snip.  Not relavent to this bug. ]
> 
> -- kernel-img.conf
> do_symlinks = yes
> relative_links = yes
> do_bootloader = no
> do_bootfloppy = no
> do_initrd = yes
> link_in_boot = no

woow people have so broken /etc/kernel-img.conf,
if you have lilo installed you want
do_bootloader = yes
so that lilo will be surely run by linux-image postinst.
 
thanks for feedback

-- 
maks