linaro-m-headless-tar-*.packages files don't include entry for kernel package

Bug #651350 reported by Peter Maydell
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Linaro image builds
Triaged
Medium
Unassigned

Bug Description

The packages manifest files like:
http://snapshots.linaro.org/10.11-daily/linaro-headless/20100929/0/images/tar/linaro-m-headless-tar-20100929-0.packages
ought to list all the packages that have gone into the tarball. However they seem to be missing the linux-image-2.6.35-1006-linaro-omap package which provides the /boot/vmlinuz-2.6.35-1006-linaro-omap and similar files. (Perhaps other packages are also missing from the list.)

(Loic suggested that this was the best place to report this bug.)

Steve Langasek (vorlon)
Changed in linaro-images:
status: New → Triaged
importance: Undecided → Medium
Revision history for this message
Steve Langasek (vorlon) wrote :

So /usr/share/live-helper/helpers/binary_manifest chroots to ./chroot and calls 'dpkg -l' to generate this list; but the chroot is consistently missing all of the packages specified in LH_PACKAGES, and contains only the bootstrap package set. This is certainly a bug, but I can't tell at a glance what the right behavior is - is the package install *supposed* to be done in the chroot directory, or is the binary_manifest supposed to chroot to ./binary instead?

Revision history for this message
Loïc Minier (lool) wrote :

We don't include kernel packages in the rootfs anymore; does this affect other packages?

Perhaps we should discuss this in the live-helper bug tracker?

Revision history for this message
Steve Langasek (vorlon) wrote : Re: [Bug 651350] Re: linaro-m-headless-tar-*.packages files don't include entry for kernel package

On Wed, Jan 26, 2011 at 11:32:15AM -0000, Loïc Minier wrote:
> We don't include kernel packages in the rootfs anymore; does this affect
> other packages?

> Perhaps we should discuss this in the live-helper bug tracker?

The live-helper version we're now using for linaro images generates a proper
manifest, but the implementation is broken and specific to the image types
we're building (i.e., not mergeable with the live-helper Canonical OEM
uses). We should track that for live-helper somehow; no opinion if
reassigning this bug is the right way to do that.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.