On Tue, 24 Aug 2010 22:57:46 -0500
Nathan Zachary <nathanzachary@gentoo.org> wrote:
> I don't think that the documentation changes should be a determining
> factor in switching to OpenRC. If we are going to endorse using OpenRC,
> the more relevant issues are the ones regarding its future development.
> The documentation can be updated in due time. Of course, that's just my
> opinion.

It's not really a determining factor in whether or not we adopt it as our default system. It's just one of the big tasks to complete if we do. I'm not arguing against using OpenRC just because the docs will require significant rewrites.

On 08/24/2010 11:57 PM, Nathan Zachary wrote:

If we are going to endorse using OpenRC,
the more relevant issues are the ones regarding its future development.


Is the future development of OpenRC more problematic than the future
development of baselayout-1? As far as I can tell, baselayout-1 never
had an upstream, and never will have one.


It seems like the debate is around openrc vs systemd or whatever. I
think the debate we need to settle first is openrc vs baselayout-1.
Otherwise we're going to end up maintaining TWO different legacy init.d
systems while we spend the next few years aiming for yet another target.


Wouldn't it make more sense to clean up openrc and get it deployed, even
if in the long-term we decide to get rid of it?


Alternatively, we drop support for openrc entirely, and tell everybody
running ~arch to move to our next target or back to baselayout-1. I
don't think we want to have three targets to maintain.


Rich

On Wednesday, August 25, 2010 12:37:34 Richard Freeman wrote:
> On 08/24/2010 11:57 PM, Nathan Zachary wrote:
> > If we are going to endorse using OpenRC,
> > the more relevant issues are the ones regarding its future development.
>
> Is the future development of OpenRC more problematic than the future
> development of baselayout-1? As far as I can tell, baselayout-1 never
> had an upstream, and never will have one.

wtf are you talking about ? Gentoo was always been the upstream of it.

> It seems like the debate is around openrc vs systemd or whatever. I
> think the debate we need to settle first is openrc vs baselayout-1.
> Otherwise we're going to end up maintaining TWO different legacy init.d
> systems while we spend the next few years aiming for yet another target.

no clue what you're talking about. Gentoo wrote baselayout from scratch, and
then rewrote baselayout-2 from scratch in C to address some fundamental issues
at the time. then Roy stepped up to do a lot of the work and when he decided
to part ways from Gentoo over POSIX shell/ebuild issues, but wanted to keep
working on baselayout-2, we allowed him to do this. so he renamed the core
bits to openrc and moved the development off of Gentoo infra.

> Wouldn't it make more sense to clean up openrc and get it deployed, even
> if in the long-term we decide to get rid of it?

it's already cleaned up. this is the "squash regressions from baselayout-1
and make sure all stable packages are happy with it" phase.
-mike

On 08/25/2010 03:06 PM, Mike Frysinger wrote:

On Wednesday, August 25, 2010 12:37:34 Richard Freeman wrote:

On 08/24/2010 11:57 PM, Nathan Zachary wrote:

If we are going to endorse using OpenRC,
the more relevant issues are the ones regarding its future development.


Is the future development of OpenRC more problematic than the future
development of baselayout-1? As far as I can tell, baselayout-1 never
had an upstream, and never will have one.


wtf are you talking about ? Gentoo was always been the upstream of it.



Uh, that was essentially my point...

Clearly upstream support is not an issue that distinguishes openrc from
baselayout-1.



Wouldn't it make more sense to clean up openrc and get it deployed, even
if in the long-term we decide to get rid of it?


it's already cleaned up. this is the "squash regressions from baselayout-1
and make sure all stable packages are happy with it" phase.


And my point was essentially that we should finish doing that, and not
bag the whole project because of the OpenRC upstream issues. Sure, we
can think about the next great thing that is coming along, but let's not
abandon the work done so far, because doing so means living with
baselayout-1 for another few more years.


I was just being a bit subtle in my argument...

Rich

Richard Freeman posted on Wed, 25 Aug 2010 16:16:29 -0400 as excerpted:

> And my point was essentially that we should finish doing that, and not
> bag the whole project because of the OpenRC upstream issues. Sure, we
> can think about the next great thing that is coming along, but let's not
> abandon the work done so far, because doing so means living with
> baselayout-1 for another few more years.

AFAIK, you're arguing the (possible, but really never became more than a
potential) debate of several months ago. As OpenRC was originally from
Gentoo's baselayout, it's not a big problem to re-adopt it as upstream
once again, certainly less of a problem at this late date than staying on
baselayout-1 stable for another several years would be likely to be, given
how legacy it is, and how close to stable it already is. The loss of
external upstream was just one more hiccup of a number of them over the
years, and isn't a big problem, especially when someone's already stepped
for the job.

Perhaps we'll eventually switch to something else, but having seen the
pains openrc went thru, I'd certainly not want to jump on to upstart or
the like at this point. Let the new round of candidates mature a bit, and
then do an evaluation. Meanwhile, what few bugs remain for openrc
stabilization pale in comparison to the bugs and adaption issues we'd have
moving to something else, and baselayout-1 really /is/ anachronistic and
not a particularly viable option at this point, so for the medium term,
openrc remains the only really viable option.

But that was pretty much decided some time ago, based on my following of
the relevant discussions here and elsewhere, so why are you arguing a
point that's not being argued any more? I believe that's what Mike's
WTFing about. It's not that you're wrong, you're not, it's that you're
debating a question that's no longer being asked.

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman

On 08/25/2010 08:29 PM, Duncan wrote:

But that was pretty much decided some time ago, based on my following of
the relevant discussions here and elsewhere, so why are you arguing a
point that's not being argued any more? I believe that's what Mike's
WTFing about. It's not that you're wrong, you're not, it's that you're
debating a question that's no longer being asked.



See:
http://thread.gmane.org/gmane.linux.gentoo.devel/67098
and
http://thread.gmane.org/gmane.linux.gentoo.devel/67098

The post I replied to cited upstream issues as a reason not to adopt
OpenRC. My point is that upstream issues are not a distinguishing
factor between OpenRC and baselayout-1.


It isn't like I'm re-opening a thread from months ago. I'm replying
directly to a point others have raised.


If there is no debate about whether OpenRC should be adopting it, then
why is it even being discussed in this way? Let's just do it...


Rich

On Thursday, August 26, 2010 13:02:15 Richard Freeman wrote:
> If there is no debate about whether OpenRC should be adopting it, then
> why is it even being discussed in this way? Let's just do it...

there is no debate. people saw Roy moving on and got scared. as i said
originally, it makes no difference to us. we're moving to openrc and it will
continue to be our default init system for the foreseeable future and the
support channel is unchanged: go to bugs.gentoo.org.

as for people who want to move to the latest shiny init, as i also said
before, nothing is stopping them from getting it working today. we've had
alternative init systems in the past that drop-in replace baselayout/openrc
and there will continue to be ones in the future.

however, until someone actually does the work to get one of the alternatives
in the tree and actually working with other packages, there is no debate to be
had as to the default init package.

i'd also highlight that openrc focuses on one thing: it exists to boot the
system and manage daemons via init scripts. it does not do all of the
extended things that systemd is taking over (inetd, crond, udevd, etc...).
-mike

Mike Frysinger posted on Thu, 26 Aug 2010 14:29:42 -0400 as excerpted:

> On Thursday, August 26, 2010 13:02:15 Richard Freeman wrote:
>> If there is no debate about whether OpenRC should be adopting it, then
>> why is it even being discussed in this way? Let's just do it...
>
> people saw Roy moving on and got scared.

That sums it up nicely, Vapier. Thanks. =:^)

--
Duncan - List replies preferred. No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master." Richard Stallman