On Tue, May 05, 2020 at 06:56:02AM +0200, Marc Haber wrote:
> On Mon, May 04, 2020 at 12:23:18PM +0100, Colin Watson wrote:
> > This sort of thing *really* needs to go upstream first, otherwise we can
> > easily end up with even more catastrophically complicated compatibility
> > code. I've commented in the merge proposal to that effect.
>
> Did you look at the diff between Debian's 10_linux and upstrem's
> 10_linux recently?
I'm familiar with it, but:
(1) that's not by itself a free pass to add more
(2) when I said "even more catastrophically complicated", that was
specifically pointing out that we already have a lot of additions
and that that's a problem
(3) several of those changes have been through specific efforts to try
to make them upstreamable, and some have been sent upstream, so
it's not just a blithe free-for-all even if the situation is more
complex than I'd like
(4) most of those changes don't add new external interfaces in the
sense of new variables in /etc/default/grub
(GRUB_CMDLINE_LINUX_RECOVERY and GRUB_RECOVERY_TITLE are
exceptions, it's true, but they're much less complex than what you
propose - for example they don't involve re-sourcing
/etc/default/grub*, which is probably an unworkable implementation
given the strange things that can be done in /etc/default/grub.d/)
(5) sarcasm isn't a persuasive line of argument
--
Colin Watson [cjwat...@debian.org]
