Bug#358397: initramfs-tools initrd.img update
Hi maks,
On Fri, Aug 04, 2006 at 04:18:14PM +0200, maximilian attems wrote:
> rechecking the issues that were reported on #358397 and #362064
> 6) debconf note
> a preinst warning of the update.
See Manoj's reply for concerns about note vs. question here. I also don't
see any reason why the note would need to happen in the preinst, which would
require a pre-dependency on debconf; it's sufficient to stop the action
prior to completion of the postinst, it doesn't need to happen prior to
unpacking the package.
> 7) races in hook install order
> no convincing backup for those, initramfs stable through the runs
I'm not sure what you mean by this. The evidence of race condition problems
comes from reasoning about the design of the hook script interface and
showing that it *can* happen (and is therefore a bug in the design), not
that it *does* happen in any particular case or on any particular upgrade
run.
> concerning 8 there is only udev that _needs_ config
> out of /etc/ the other hook scripts have no needs there.
> that is an open bug in mkinitramfs. it shouldn't run
> unless the shipped /etc/initramfs-tools/initramfs.conf
> is unpacked. (udev postinst update_initramfs() checks
> explicitly for that condition, but that needs to be done
> for all hooks in mkinitramfs.) mkinitramfs relies on the
> existence of that conffile.
> i will fix point 8 and add a warning for 1,
> in next initramfs-tools upload, once incoming version is
> in etch.
That sounds like an improvement to me, and from what I can see/remember,
ought to be sufficient IMHO to resolve the RC elements of this report. I
can't say I'm altogether satisfied with the idea of initramfs images being
auto-updated, but I can't see any reason left to consider them RC if
adequate precautions are taken to prevent working images from being
overwritten with broken ones.
--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
vorlon@debian.org http://www.debian.org/
Reply to: