Hello, On 31 December 2015 at 15:40, Iain R. Learmonth <[email protected]> wrote: > Hi, > > On Thu, Dec 31, 2015 at 11:20:44AM +1100, Michael . wrote: >> Iain, many people want a stable release on newer hardware so they do >> need a newer kernel to support the hardware. Your testing release does >> not fit all use cases and the suggestion made by Mutlu is worthwhile >> and if I may just add it is achievable in live-build. > > As you say, it is achievable with live-build, so for those that need this > image they can create one themselves. > > For every image produced there is extra workload for creating the > configurations and for the testing of those images to ensure they're working > correctly. > > One option I would be happy to explore for stretch would be to include the > kernel from stable-backports in the stable images alongside the stable > kernel, and you can choose to use the newer kernel from the boot menu if you > wish, although this will need discussion as it would mean adding the > stable-backports repository to the stable images and we would need to be > sure that this wouldn't have any negative effects.
I build such images. I use patches for live-build that allow including multiple kernels without the live-build part that generates the syslinux menus choking. You can use pinning to avoid installing from backports by default and a hook script to install any number of kernels you like. In fact I install kernels from sid but installing the ones from backports would work the same way. I can post the patches somewhere if there is interest in including them. They are against the shell script and were not accepted because that part of live-build was supposed to be rewritten in python. Thanks Michal
