Re: Bug#863801: grub-coreboot: fails to upgrade from jessie to stretch if init-select was installed
On Sat, Jun 24, 2017 at 07:01:11PM +0200, Michael Biebl wrote:
> Am 24.06.2017 um 17:01 schrieb Simon McVittie:
> > That doesn't solve the problem of the obsolete conffile breaking grub,
> > though.
>
> Indeed not. But it answers the question whether init-select should be
> NMUed in unstable.
I agree that that part of it is now moot.
> > Should the grub maintainers edit the conffile in-place as
> > suggested (a Policy violation), or delete it or move it out of the way
> > (also a Policy violation), or is there some other escape route possible
> > here?
> >
> > It occurs to me that asking the CTTE for advice might be useful: they'd
> > probably find it a refreshing change to have a question that is not a
> > request to choose one side of a heated dispute between developers :-)
>
> Since it's pretty obvious that init-select is supposed to be removed, I
> wouldn't have a problem with simply forcefully removing the offending
> init-select.cfg conffile (and it's probably safe to drop this migration
> code after one release cycle.
>
> Asking the CTTE certainly doesn't hurt.
I think I probably agree that removing the file is sensible, but my
experience is that it's a good idea to have an unusual amount of review
of unusual and difficult-to-undo conffile handling ideas: so I've
written this up for the TC's consideration as
https://bugs.debian.org/865929.
Thanks,
--
Colin Watson [cjwatson@debian.org]
Reply to: