On Mon, May 16, 2016 at 09:27:43PM -0400, Steve Litt wrote:
> > I sincerely wish the remedy of installing openrc and pinning systemd
> > would work, but I'm sure it won't on the mid-long term.
>
> It obviously can't. The same Debian leadership who voted "no" (or was
> it "we don't need no steenkin GR") on supporting multiple inits now
> determines whether the OpenRC package is maintained or not. That's no
> foundation on which to build init choice freedom.

Thoroughly no! Please don't spread such misconceptions -- it's voicing such
views that brings the most ammunition to the pro-systemd faction.

Openrc is undermaintained in Debian not because of some cabal, but because
of manpower issues: no one does real work on it currently, and its listed
maintainers are hardly active.

True, systemd-related changes (like, in init-system-helpers) do add to this
burden, but that's not different from, say, burden on libpng using packages
caused by the png12->png16 migration. This is a good part of maintainer's
duties. You can't really blame the additional work for failures when even
the basic work isn't done.

I admit to be a part of the problem -- I should have stepped in to help
myself. My excuses here are 1. not knowing much about init/rc systems, and
2. already having my hands in too many baskets, but I guess most of us can
say something of this kind. End result is that there's no one to really
maintain openrc. And sysvinit, for that matter.


The "Debian leadership" has pretty little power: the DPL manages finances,
public relations and delegations (in practice). The CTTE is hardly ever
invoked, and their rulings on the systemd issue can't be considered evil:
* "systemd shall be the default for jessie (and by extension stretch)" --
well, they did have to choose something. That we are unhappy with the
choice doesn't mean they're wrong in making it.
* "support for non-default init systems should be kept" -- isn't this what
we want?

The source of evil is the near-absolute power maintainers have on their
packages. Other than the really unwieldy stick of complaining to CTTE,
there's no real way to fix issues directly. For example:
* the d-i team refused a well-tested patch fixing an obvious (as in, it
doesn't work as documented) bug on debootstrap --exclude that would enable
us to install sane inits without clumsy and bug-prone schemes that install
systemd then remove it
* the utopia team keeps breaking their packages in non-systemd scenarios
(non-systemd Linux, kfreebsd). Both in jessie and current unstable you
can't shutdown/reboot/suspend/hibernate from GUIs, etc.

> By the way Jaromil, several days ago I told Rick the exact same stuff
> you wrote here. If it were only systemd, and if the Debian DDs (boy do
> I hate that phrase) were willing to admit their mistake, we wouldn't be
> here. It's about Debian no longer being safe stewards of GNU/Linux.

The DDs are a diverse group, there's no "their mistake". You may disagree
with the actions of certain individuals, and thus the way package X is
maintained, but other than a few cases (like d-i and utopia) this is
entirely because of no one bothering to do the work to make alternatives
viable. Like, xfce has currently no non-pulseaudio volume control not
because their maintainers are evil but because no one wrote an ALSA backend
when gstreamer dropped volume control APIs (considering them out of scope).
All it takes to fix this is to actually write the code.

Disclaimer:
[ -z "$(gpg --keyring /usr/share/keyrings/debian-keyring.gpg \
--no-default-keyring --list-keys kilobyte@???)" ]
returns false.

--
A tit a day keeps the vet away.