On Wed, Aug 31, 2011 at 11:44 AM, Mark Phippard <markphip_at_gmail.com> wrote:
> On Wed, Aug 31, 2011 at 12:31 PM, Hyrum K Wright <hyrum.wright_at_wandisco.com>
> wrote:
>
>>
>> As for a specific date, I'm hesitant to commit to something. While I
>> certainly appreciate the utility, my crystal ball just isn't that
>> clear. I guess it boils down to the fact that I've been telling
>> people my guess about 1.7's release date for *years*, and I'd rather
>> under promise and over deliver on this one than vice versa. Maybe I'm
>> getting commitment fatigue. :)
>
> I am not asking you to predict whether this will be our final release
> candidate or not. I am saying let's publish a date where we will start the
> process and do the release if this RC holds up. If we have to re-soak for
> some reason then obviously we have to reset the date. I would also expect
> you to look at your own schedule. For example the 4-week date might fall on
> a time when you are traveling so that it would have to be done the next
> week. This is the sort of think the rest of us cannot know if we just try
> to look at the calendar and add 4 weeks. It does not do any good for me to
> project the date if you already know that you are not going to be able to
> actually do the work on those days.
I think it's safe to say that I'm generally available, and--aside from
my Sunday commitments mentioned earlier--make a best effort to handle
release matters within a few hours, even when traveling or otherwise
occupied.
>> You mention being flexible with the date if something comes up. In
>> which case, why even set a date at all? It feels a lot like releasing
>> a release candidate that isn't really a candidate for release. ;)
>>
> I would expect us to make a best-effort commitment. If you say it is going
> to be on a specific date I would expect you to do it on that date unless
> something comes up. Like I said, obviously if the software itself is not
> ready we have to reset the date. Otherwise, I am just saying that if
> something you do not foresee were to come up. Such as needing to do
> something work/school related then I would understand if you have to notify
> us of a change.
That's just the thing: if we reserve ourselves that flexibility, then
it makes the date worthless, so why even bother setting one?
> I want to know because there is a lot of stuff I have to do when a release
> comes out and I have to coordinate the resources I need to do it all. Since
> some of those resources are shared, it even impacts non-SVN related
> projects. I also want to do a GA Subclipse release right around the GA
> date. We have already done the 1.7 compat work, but knowing the date gives
> an idea how much time we have to work on new features before we freeze for
> the final release. I am sure I am not the only one in this boat. I assume
> there are people at WanDisco that want to know when they will have to build
> and QA binaries for the release.
The project has always had a "release when it's ready" mind frame, and
I think committing to a firm date undermines that. Coordinating with
downstream is important, certainly, but we delude ourselves if we
think that people are trying to schedule launch parties around the
release of Subversion 1.7.0.
> Is there any reason we cannot simply be public about the forecasted date if
> this RC holds up?
This mailing list *is* public. If folks are really interested, they
(or their lackeys) can hit up the archives and perform their own
deductions. We aren't having these discussions in a closet.
For what it's worth, the text in the release announcement reads thus:
This release candidate begins the four-week "soak" period to allow for
further testing, and barring show-stopping bugs, the final 1.7.0
release can be expected on or near Sept. 28.
That's about as committal as I feel comfortable with at this point.
-Hyrum
--
uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com/
Received on 2011-08-31 19:21:19 CEST