On Tue, Aug 23, 2005 at 02:10:08AM +0300, Cyril Plisko wrote:
> On 8/23/05, Stephen Lau <[EMAIL PROTECTED]> wrote:
> > On Mon, Aug 22, 2005 at 03:12:23PM -0700, Mike Kupfer wrote:
> > > >>>>> "Cyril" == Cyril Plisko <[EMAIL PROTECTED]> writes:
> > >
> > > Cyril> On 8/23/05, Stephen Lau <[EMAIL PROTECTED]> wrote:
> > >
> > > >> Of course, our long term plan is to just move the daily source-gate
> > > >> out externally, but this will take a little bit more time.
> > >
> > > Cyril> "little bit" sounds very encouraging :) Can you please set some
> > > Cyril> expectations ? Are we talking about 1-2 more builds, another
> > > Cyril> couple of months or longer than that ?
> > >
> > > My guess is more than a couple months, though I don't know how much
> > > longer. Nightly updates require more automation than we currently have
> > > in place. And that requires some changes to our internal source
> > > organization. I'm working on that, but I've still got a fair amount of
> > > work left.
> > >
> > > mike
> >
> > Yeah.. what mike said. :-)
> >
> > Sorry, I didn't mean "little bit" literally; I didn't mean to get your
> > hopes up prematurely. I just wanted to make sure people knew what we
> > were working towards.. :)
>
> Cool !
> Thanks for sharing that with the rest of the community.
> Do you think it is appropriate to [publicly] discuss what the
> problems are with the making that (nightly updates) happen ?
> I certainly think that community will benefit from it.
Well, there are two goals that I see here:
1) Nightly source drop updates, while the gate is still within Sun
This is the closer objective, and is basically the same thing as
what Mike, Keith, and I have been doing - except on a daily frequency
rather than a monthly or bi-weekly frequency. The main problem here is
that not every step of a source drop can be automated fully. The main
problem is that it's hard to take "snapshots" of the Nevada source,
strip out the closed stuff and patch things like Makefiles in an
automated manner... so some of this work has to be done by hand by Mike
or myself. Doing this on a daily frequency can be tedious and
time-consuming, if not impossible given the time to build/test each
drop.
2) Move the Nevada gate outside of Sun
This is the ideal (and long term) goal, but there are
higher-level problems (some policy, some technical) such as how to
manage external projects. We need some infrastructure to allow people
(by people I mean external contributors) to create workspaces,
bringover, putback, merge, etc. etc. So, part of this is the age
old problem of source code management. There are probably other
aspects to this that Mike or Keith are more familiar with...
cheers,
steve
--
stephen lau // [EMAIL PROTECTED] | 650.786.0845 | http://whacked.net
opensolaris // solaris kernel development
_______________________________________________
opensolaris-help mailing list
[email protected]
