> -----Original Message-----
> From: Hyrum K. Wright [mailto:hyrum_wright_at_mail.utexas.edu]
> Sent: vrijdag 12 februari 2010 20:57
> To: Subversion Development
> Subject: Setting a 1.7 branch goal
>
> I've been visiting with a number of users and other interested parties
> about the future of Subversion, but one of the things I'm constantly
> hearing is "when will 1.7 be released?" It'd be nice to be able to
> address that. :)
>
> 1.7 will be an important release, but it feels as if we've been
> stalling a bit lately. I'm excited to see Mike, Julian and other
> getting involved, and I think the work we've left to do is well-scoped
> (if not well understood by everybody yet).
>
> In the past, we've been hesitant to commit to a specific date, but I
> think making such a commitment to ourselves would be both useful and
> motivating. To kick things off, I propose the following:
> * We shoot for a mid-May branch date, say May 15.
> * We can use the June Berlin hack-a-thon to do a final push for any
> bugs which have popped up during the RC phase (or a push toward an RC
> if something postpones the branch date).
> * A final release happens sometime mid-summer
I don't think we should set a specific date unless we are willing to cut
features at that date.
At this time we would have to cut WC-NG as it misses certain stability
guarantees.
If you press ^C somewhere during update you can now have half updates on
files (the data, but not the conflict or its properties, etc.). WC-NG must
be 100% stable or not part of the release.
And if we would cut WC-NG, we would have to cut most of the libsvn_client
stuff too as we implemented svn patch and other features on top of WC-NG.
>
> I'd be very happy if this process were to accelerate, but I think as
> things look right now, the above looks reasonable.
On Subconf 2009 we had an indicative calculation of getting to a single
administrative area around January and working copy format stable early
after that; Beta around April and release around June.
I don't see us getting to a single administrative area in the next few
weeks.. so just hoping for a beta around may would be a gigantic increase in
development speed while no major milestones have been reached since October.
(The in-DB properties code was working then, but still not in use as primary
property storage)
And before beta we have to get a stable working copy format. We can't ask
beta users to expect their working copies to be incompatible broken. (Of
course we can still change it if we find design bugs, but we can't go into
beta with known issues that have to resolved later)
Getting a group of new developers focusing on WC-NG will only get us an
acceleration when these developers already have all the knowledge required.
Providing that knowledge to them causes an initial slowdown. (Read any
software project management book about this)
Yes, I would like to speed up the development, but just throwing a date on
the table won't help :(
I wish I could spend more time on WC-NG but my $dayjob doesn't sponsor
development work on Subversion and free time is spare with a baby in the
house.
And the Windows buildbot out of order delays starting up development every
time as there are still unresolved Windows test failures..
Bert
Received on 2010-02-13 10:47:25 CET