:: Re: [devuan-dev] Sharing the love
Góra strony
Delete this message
Reply to this message
Autor: Daniel Reurich
Data:  
Dla: KatolaZ, devuan developers internal list
Temat: Re: [devuan-dev] Sharing the love
On 19/06/18 17:40, KatolaZ wrote:
> On Tue, Jun 19, 2018 at 09:34:50AM +1200, Daniel Reurich wrote:
>
> [cut]
>
>>> Let me say that what we really need now is not the sort of "package
>>> maintainer" clergy found in Debian, but developers with far fetched
>>> vision including sustainable plans and capable of leveraging Devuan's
>>> adoption into segment of private and public sector infrastructure,
>>> since it may be an high profitable game to assume liability for a
>>> system we know it is simple and easy to manage. Also we do need all of
>>> you and especially those who are thinking of deep strategies to avoid
>>> entaglement with the least possible effort. With Beowulf there are
>>> possibilities at the horizon that may reduce drastically the number of
>>> packages we need to fork, by means of developing a proper systemd shim
>>> for instance.
>>>
>> I think that you'll find a systemd shim will only be a temporary patch,
>> perhaps necessary in the short term, but one that eventually hold us back.
>>
>> What we really need is to build sufficient maintainer capacity and
>> capability to fork more packages and excise this systemd cancer once and
>> for all. Beating it into a corner where it can be managed will only be
>> a short term solution. If we don't slay the beast, it will continue to
>> grow and infect more components ever more invasively, and with Debian
>> developers actively removing sysvinit scripts etc, we will end up with
>> lots of work arounds that will become more of a burden to maintain then
>> cleanly forked packages.
>>
>> After all wasn't Devuan supposed to be a "Fork of Debian" and not just a
>> derivative relying on Debian's good graces??
>>
>
>
> Wishing is cheap, making it a reality is another story :) We don't
> have the energy to fork more packages at the moment, so we need a
> solution that provides a path of least resistance. I really hope we
> will have dozens and dozens of maintainers in the future, but at the
> moment we simply don't have them. A maintainer is a person who commits
> time to take care of a package at least for an entire release
> cycle. The best we have seen so far is many people who remove
> libsystemd0 from the dependencies of one package, and then disappear.


Indeed, but it's not merely wishing. It is a fore gone conclusion that
we will have to fork and maintain more packages because we have already
seen some debian maintainers actively removing sysvinit scripts from
their packages seemingly in reaction to the existence of Devuan.

Right from the beginning, it was the expectation that we would move to
be more independent from debian and diverging from them when it was
either necessary or provided significant benefits.
>
> I would humbly point out that most of the packages forked for jessie
> by several "maintainers", in an impetus of enthusiasm, had to be
> adopted either by me or by parazyd for ASCII. I am not pointing
> fingers against anybody: I am just stressing what the reality is,
> compared with what we would have liked it to be.


That's noted, and I know I'm responsible for a fair few of those. Part
of the problem is that we are working backward from stable, when we
should be working from experimental and unstable forward. Another part
of the issue is we don't have any notification that shows pending
changes to packages we've forked. Currently we're relying on people
filing bug reports.
>
> Let's discuss through this stuff, but please, let's keep our feet on
> the ground, and remember how much of an effort was required by all of
> us to push ASCII out of the door with the current number of forked
> packages (which is about 1.5% of all the Debian packages...).
>

Sure, but let's neither stick our heads in the sand and pretend that
some maintainers and developers in Debian will not continue to react
badly to the advent and growing recognition of Devuan, or that others
will perhaps be looking to abandon Debian and join Devuan instead. At
some point in the next release or so we need to have the capability to
maintain a fork of many more packages with a security team as well.

I expect for Beowulf +1 we will need to have forked somewhere between 3%
and 5% of packages - in fact we should aim for doing that. I am aware
this puts some stress on manpower and resources, but that is a good
thing. There is no time for sitting on the fence.



> My2Cents
>
> KatolaZ
>
>
>
> _______________________________________________
> devuan-dev internal mailing list
> devuan-dev@???
> https://mailinglists.dyne.org/cgi-bin/mailman/listinfo/devuan-dev
>



--
Daniel Reurich
Centurion Computer Technology (2005) Ltd.
021 797 722