Hello all,
Time to put on my Project Lead hat, I suppose.
On 25/07/2020 14:12, Fungi4All wrote:
Thank you very much for the clarifications Lee, I am sure many will
find an interest in the content as well.
I only use openbox and borrow some desktop applications here and
there from other desktop sets, like lxterminal and lxappearance.
Some I have to admit instead of building on my own I borrowed from
void-musl. So much of the ongoing change doesn't really affect me.
If you only offered i3 and jwm that would have been fine and plenty
for my needs, but I understand others prefer a more complete/full
desktop. The only sources of information about sddm and ck problems I
have found relate to setup and possibly the existence of a .xinitrc
calling up ck. Some DMs seem to have an internal dependency for
ck-launch ck-daemon and if an auxiliary file/conf calls up ck from
within a ck session that breaks ck. But I haven't had direct
experience with sddm for a long while.
We do have SDDM mostly working with ConsoleKit, but there are bugs and
the patches are not merged upstream. And upstream isn't interested, so
each bump we have to reintegrate the patches, which means more work.
The main issue is that the ConsoleKit daemon crashes after 10 minutes
idle if media is playing. This is reproduceable by not just us, but
also OpenBSD. Once the daemon crashes, you are no longer able to reboot
or shut down the computer without using a root command prompt.
For a distro that wants to be user-friendly, or at the very least *not
user hostile*, that is completely unacceptable.
And if the OpenBSD people can't even figure it out, there is little hope
for us, without a significant amount of time we could be spending on
better things (like below).
I would rather see development and energy spent on the core system,
init/service management than getting a screenlocker working with
live-video-feed. I assume there are reasons to focus some marketing
efforts one direction or another and it is not really my business.
I wish I had the abilities to contribute to the project but I am
still in the learning process to reach stage 0.
The core system for 1.0 is done. Now we are focusing on the polish so
people without much Linux experience can use it, so that we can ship 1.0
and start looking to the future.
We are still interested in porting the NetBSD userland to replace
coreutils, and replacing libtool with slibtool, and various other de-GNU
stuff. We are still interested in doing the plumbing work to make
language runtimes run on ppc64/musl. We have aarch64_be, armv6/armv7,
and SPARC64 ports planned.
But the core system is done for 1.0. And now it's time to polish it.
And I would rather work on those goals than try to keep CK2 chugging
along, just so we can avoid shipping elogind.
I want to note here as well that I am working with actual real people
locally (over the phone, due to the global plague) to understand what
real people need.
I am working with young and old. There is everyone from a recovering
cancer patient, to a grandmother, to a photographer, to a shop owner.
(And yes, those are all different people.) Nobody has expressed an
opinion on elogind, but plenty have expressed opinions on wanting to be
able to choose "Shut Down" from Plasma's menu and have it work reliably.
At the end of the day, Adélie has a service to people. In fact, we are
a Public Benefit Corporation in the State of Oklahoma with a specific
expressed purpose in our legal documents of enriching the lives of
people who rely on technology by providing them a free (as in beer, and
as in libre) operating environment that is reliable, secure, and stable.
That overrules politics every time, even when I don't like it.
And I don't like elogind. But I like it better than the alternatives.
It is not that I am leaving, it is more that I can not promote
Adelie
as much as I did. I like Kiss (
k1ss.org) and ataraxia projects as
well but Adelie had more intersection between what I at least wanted
and expected.
Ataraxia switched to elogind in April. This was part of what informed
our decision to really give up on CK2; even the staunchest distributions
can't keep that pile of poop running any more. Which is sad, but nobody
wants to touch the CK2 code. I gave it a good look and it is possibly
worse than elogind inside.
Someone on our IRC said that there isn't much difference anyway, since
CK2 and elogind implement mostly the same APIs (CK2 is also D-Bus and
coined the 'login1' name) and are both
freedesktop.org projects.
I have barely gotten up by being slapped by Void's move to
elogind,
and now I get hit again :) Just when I had gotten s6 and 66 working
on both and keeping up with development, now I have to go try to
remove busybox from K1ss and try my luck there.
They have their own issues. No NLS support, so English only, because
they won't consider gettext-tiny as a replacement for gettext. Their
choice of default media player has an erratic upstream that has been
removing and changing code without explanation. etc etc.
But I will admit their system is very flexible, and they have a good
alternatives system. We need to work our own alternatives system out.
At least you haven't adopted .zstd compression, so there are
still
some score points for Adelie.
I don't see us putting anything from FB in core, similar to how we won't
ship googleware, though for different reasons. (FB has a track record
of less than great security; googleware is just written poorly, not
portable, and gross.)
Best,
--arw
--
A. Wilcox (awilfox)
Project Lead, Adélie Linux
https://www.adelielinux.org