Skip to Content.
Sympa Menu

sm-discuss - Re: [SM-Discuss] Let's make Source Mage not suck!

sm-discuss AT lists.ibiblio.org

Subject: Public SourceMage Discussion List

List archive

Chronological Thread  
  • From: Jeremy Blosser <jblosser-smgl AT firinn.org>
  • To: sm-discuss AT lists.ibiblio.org
  • Subject: Re: [SM-Discuss] Let's make Source Mage not suck!
  • Date: Thu, 4 Jan 2007 23:54:31 -0600

On Jan 04, Daniel Goller [dgoller AT satx.rr.com] wrote:
> On Thu, 04 Jan 2007 13:38:14 -0800
> Andrew <afrayedknot AT thefrayedknot.armory.com> wrote:
>
>
> > >
> > > Perhaps a trial of the freeze method is in order?
> > >
> >
> > Well, stable-rc kind of approximates that freeze, we dont put version
> > bumps in it, and, in theory we fix the bugs in the spells we care about.
> >
> > Hypothetically, we could remove stable-rc and freeze test, then once
> > test meets our standards (some subset of spells "works"), then release
> > it. However people will just start submitting there version bumps to
> > devel (where else?). And devel will become what test is now. Devel right
> > now means something else, so we'd end up merging the current definition
> > of test with that of devel. Or, we'd make another "really-devel" grimoire.
> >
> > The thing about stable-rc is that it recognizes there are 4 stages,
> > not 3. If we remove one of the steps, we inadvertantly merge stages
> > together, and then, as history shows, we'll waste yet more time arguing
> > about what the 2 stages represent. So, I think freezing is a step
> > backwards.
> >
> >
> >
> > -Andrew
> >
>
> So in theory we have test and stable.
> When it is time for a freeze, test gets snapshot into stable-rc.
> Once stable-rc has no more bugs against it, it becomes 'stable'.
s/no more bugs/no more bugs identified as "must fix"/
> At which point stable-rc has no need to exist.
> So the freeze everyone talks about is stable-rc, or should be.

100% correct with the noted correction. This is how it was set up to work,
and we've done at least the last two stable releases by this methodology.

> This leaves test open for those who want to provide versions bumps while
> stable-rc is "readied". There are no version bumps in stable-rc, unless
> perhaps security related items.

Yes, or if a version bump is the right way to fix a must fix bug and it
doesn't introduce other changes that would need their own regression
testing. Minimal changes to fix only the noted issues is the preference,
regardless.

> And it allows stable-rc to become the new 'stable' by fixing it while
> 'stable' users have something to use, w/o having to worry about anything
> being dropped on them, that has not gotten proper testing.

Yes.

> We don't need to freeze test to have a freeze, let's face it, the idea is
> "if we freeze test, people have no choice but to work on stable-rc",
> which some will welcome, and others will not (i think at least we will
> have both sides).

Yeah, if we do this we'll just have people start working on another branch
as test or quit altogether. We can't force people to work on what we want
them to, not around here.

> What we need is a motivation for people to work on stable-rc, not a way
> to force them to do it.

Speaking only for myself, it's not motivation for me, it's just time. This
is still my top volunteer project, I just haven't had hours after work+kids
lately for any volunteer stuff.

For at least some other people it does seem to be motivation I guess.
People say they want a distro with frequent stable releases but often work
on the "cool side projects" instead of the routine stuff needed to get
stable releases out. Not that I really blame them for wanting to work on
the cool stuff, I just think people need to recognize the dichotomy between
what they say they want and what they work on.

No I don't mean anyone in particular, I couldn't name names there if I
tried, it's just an impression. I also know most of our people with time
do tend to get cracking on the routine stuff when anyone takes the lead to
get it organized, but we seem to have a limited group who take that lead,
and they often suffer from lack of time.

> This is how i see it, let's treat stable-rc as the release candidate it
> is, not turn test into one, nor use freezing of test as a 'motivator', i
> think it would backfire/have no effect.
> Leaving test open allows people to do what they perceive as day to day work.
>
> No, i have no one size fits all suggestion on how to motivate people.
> Can't say my motivation is at a peak here either.
>
> Daniel

Attachment: pgpdfP1tmPKNj.pgp
Description: PGP signature




Archive powered by MHonArc 2.6.24.

Top of Page