Re: [gentoo-dev] Re: [gentoo-dev-announce] OpenRc-0.12 is coming soon

[Rich Freeman]

On Sat, Aug 3, 2013 at 6:03 AM, Markos Chandras <hwoarang@g.o> wrote:
> On Aug 3, 2013 10:06 AM, "Donnie Berkholz" <dberkholz@g.o> wrote:
>> On 15:36 Fri 02 Aug     , William Hubbs wrote:
>> > I do not know of any breakage personally. It does work on my system, and
>> > I know of others who are using OpenRc from git successfully. Some are
>> > OpenRc team members, and at least one is a Gentoo user.
>>
>> Man, in terms of how to phrase things, this is way wrong.
>>
>> If you're comfortable with your stuff breaking really? No. If you want
>> to help improve Gentoo.
>>
> I am not comfortable with this either. If you think the new openrc will
> likely break things please mask it for a few days. Do not expect all users
> to read the mailing list.

I think the only real issue is the wording here.

He said that several are running it successfully.  I think that this
means that is sufficiently stable to add to ~arch unmasked.  Adding it
masked won't really accomplish anything unless there is a call for
volunteer testers anyway.  (By all means williamh should do so if he
feels it is prudent.)

~arch is UNSTABLE - from time to time things are expected to break.
We shouldn't be committing known breakage, but since ~arch is where
things are tested it is more likely that inadvertent problems will
sneak in.  Openrc is one of those things that is inconvenient to have
break, so the heads-up is a good idea, and perhaps some individuals
will prefer to delay updating this particular package.  The fact that
everybody will pick a different wait time also staggers the rollout so
we don't have 50,000 broken systems on day 1.

I don't expect all users to read the mailing list, but users who are
interested in testing our experimental packages probably should do so.
 If users aren't interested in testing our experimental packages, they
shouldn't be running unstable keywords - certainly not for openrc.

And yes, I fully expect a few people to chime in and point out that
they find ~arch less buggy than stable because more devs run it.  The
fact remains that accepting ~arch means that you're willing to deal
with bleeding-edge packages that have not been tested.  That is a fine
and noble thing, and maybe it works out better for you 99 times out of
100.  However, 1 time out of 100 it might be openrc that bugs out on
you.  You should always be prepared for this (and really - it isn't
that hard to run quickpkg before doing big upgrades and have a rescue
CD lying around).

Rich