First 1.7.0 prerelease: June 1

[Hyrum K Wright <hy...@hyrumwright.org>]

Fire up your testing environments: they're going to get some use over
the next couple of months.

In an effort to release early and often, I'd like to cut the first
1.7.0 pre-release on June 1.  This will likely be a beta release,
since there are still blocking issues, but I hope to get this into the
hands of users (and third-party consumers) soon-ish.

>From that point, we'll continue to release beta's until we branch,
which will be when the "1.7.0" issues are all closed.  After the
branch, we'll start the release candidates and the soak period.

Thoughts?

-Hyrum

Re: First 1.7.0 prerelease: June 1

[Stefan Sperling <st...@elego.de>]

On Wed, May 18, 2011 at 12:17:18PM +0200, Hyrum K Wright wrote:
> Fire up your testing environments: they're going to get some use over
> the next couple of months.
> 
> In an effort to release early and often, I'd like to cut the first
> 1.7.0 pre-release on June 1.  This will likely be a beta release,
> since there are still blocking issues, but I hope to get this into the

Or maybe "alpha"? But whatever, it's just a name...

> hands of users (and third-party consumers) soon-ish.
> 
> >From that point, we'll continue to release beta's until we branch,
> which will be when the "1.7.0" issues are all closed.  After the
> branch, we'll start the release candidates and the soak period.
> 
> Thoughts?

I think June 1 is good timing.

Re: First 1.7.0 prerelease: June 1

[Hyrum K Wright <hy...@hyrumwright.org>]

On Fri, Jun 3, 2011 at 2:54 AM, Stefan Fuhrmann <eq...@web.de> wrote:
> On 03.06.2011 04:16, Hyrum K Wright wrote:
>>
>> On Wed, May 18, 2011 at 5:17 AM, Hyrum K Wright<hy...@hyrumwright.org>
>>  wrote:
>>>
>>> Fire up your testing environments: they're going to get some use over
>>> the next couple of months.
>>>
>>> In an effort to release early and often, I'd like to cut the first
>>> 1.7.0 pre-release on June 1.  This will likely be a beta release,
>>> since there are still blocking issues, but I hope to get this into the
>>> hands of users (and third-party consumers) soon-ish.
>>>
>>>  From that point, we'll continue to release beta's until we branch,
>>> which will be when the "1.7.0" issues are all closed.  After the
>>> branch, we'll start the release candidates and the soak period.
>>
>> Update: the circus surrounding the 1.6.17 release has pushed this
>> timeline back just a bit.  As the buildbots are all currently green, I
>> plan to use r1130886 as the magic rev for 1.7.0-alpha1 and roll and
>> post the tarballs tomorrow morning (CDT).
>
> Thumbs up!
>
> It just occurred to me that the alpha should come
> with a big "1.7 APIs and UI are not final" warning.

That will be one of *many* warnings issued with these pre-releases. :)

-Hyrum

Re: First 1.7.0 prerelease: June 1

[Stefan Fuhrmann <eq...@web.de>]

On 03.06.2011 04:16, Hyrum K Wright wrote:
> On Wed, May 18, 2011 at 5:17 AM, Hyrum K Wright<hy...@hyrumwright.org>  wrote:
>> Fire up your testing environments: they're going to get some use over
>> the next couple of months.
>>
>> In an effort to release early and often, I'd like to cut the first
>> 1.7.0 pre-release on June 1.  This will likely be a beta release,
>> since there are still blocking issues, but I hope to get this into the
>> hands of users (and third-party consumers) soon-ish.
>>
>>  From that point, we'll continue to release beta's until we branch,
>> which will be when the "1.7.0" issues are all closed.  After the
>> branch, we'll start the release candidates and the soak period.
> Update: the circus surrounding the 1.6.17 release has pushed this
> timeline back just a bit.  As the buildbots are all currently green, I
> plan to use r1130886 as the magic rev for 1.7.0-alpha1 and roll and
> post the tarballs tomorrow morning (CDT).
Thumbs up!

It just occurred to me that the alpha should come
with a big "1.7 APIs and UI are not final" warning.

-- Stefan^2.

Re: First 1.7.0 prerelease: June 1

[Hyrum K Wright <hy...@hyrumwright.org>]

On Wed, May 18, 2011 at 5:17 AM, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> Fire up your testing environments: they're going to get some use over
> the next couple of months.
>
> In an effort to release early and often, I'd like to cut the first
> 1.7.0 pre-release on June 1.  This will likely be a beta release,
> since there are still blocking issues, but I hope to get this into the
> hands of users (and third-party consumers) soon-ish.
>
> From that point, we'll continue to release beta's until we branch,
> which will be when the "1.7.0" issues are all closed.  After the
> branch, we'll start the release candidates and the soak period.

Update: the circus surrounding the 1.6.17 release has pushed this
timeline back just a bit.  As the buildbots are all currently green, I
plan to use r1130886 as the magic rev for 1.7.0-alpha1 and roll and
post the tarballs tomorrow morning (CDT).

-Hyrum

RE: First 1.7.0 prerelease: June 1

[Bob Archer <Bo...@amsi.com>]

> Fire up your testing environments: they're going to get some use
> over
> the next couple of months.
> 
> In an effort to release early and often, I'd like to cut the first
> 1.7.0 pre-release on June 1.  This will likely be a beta release,
> since there are still blocking issues, but I hope to get this into
> the
> hands of users (and third-party consumers) soon-ish.
> 
> From that point, we'll continue to release beta's until we branch,
> which will be when the "1.7.0" issues are all closed.  After the
> branch, we'll start the release candidates and the soak period.
> 
> Thoughts?

Sweet... I for one appreciate everyone's work on 1.7 and am really looking forward to it. Thanks to everyone. While I "can't wait" figuratively for this release I actually can wait literally because I would rather have it later and done than sooner and iffy.

BOb

Re: First 1.7.0 prerelease: June 1

[Daniel Shahaf <d....@daniel.shahaf.name>]

Mark Phippard wrote on Wed, May 18, 2011 at 07:43:12 -0400:
> Give the community a date.  If it doesn't look like you can hit the
> date, then update the community with the new date and move on.  It
> seems like the shame in possibly missing a date is being given too
> much weight.  The community still wants and needs those dates.

What is the point of giving a date that doesn't mean anything?

Re: First 1.7.0 prerelease: June 1

["C. Michael Pilato" <cm...@collab.net>]

On 05/18/2011 03:35 PM, Mark Phippard wrote:
> I think there are two primary things that you could accomplish by
> being together and able to talk to each other:
> 
> 1) Talk about what stands between where we are now and a finished
> release.  Make note of any items not captured in the tracker and add
> them.
> 
> 2) Talk about some of these items in detail and the right approach to
> solving them.
> 
> Doing #1 will give people confidence that there is not much left to
> do, or it will at least reveal all the things we have not captured
> already.  Doing #2 should give a sense on how much work it will take
> and possibly who can do it.

As it turns out, your mail arrived while we were doing exactly these things.
 I think perhaps it's unfortunate that the sole bit of "group communication"
from this hackathon thus far has been this one somewhat non-committal
proposal from Hyrum.  In reality, there are plenty of discussions at various
depth levels happening in the room here.

No doubt you've seen several new issues filed -- the result of some of us
realizing that our collective Inboxes and cerebellums are quite the volatile
store for alot of information that's critical to our ability to plot and
traverse the course of this release.  And in a few minutes I'll be
committing up my notes from the agenda-dictated topics we've discussed so far.

-- 
C. Michael Pilato <cm...@collab.net>
CollabNet   <>   www.collab.net   <>   Distributed Development On Demand

Re: First 1.7.0 prerelease: June 1

[Mark Phippard <ma...@gmail.com>]

On Wed, May 18, 2011 at 8:52 AM, Greg Stein <gs...@gmail.com> wrote:

>> We are not the only software project in the world working on hard
>> problems or that finds it difficult to provide estimates.  It sucks,
>> but part of being a professional is doing the best you can.
>
> Do you think that you can plan around *volunteers*?

Yes, I do.  I assume volunteers will only sign themselves up for tasks
they believe they can complete.  In many cases, they may choose to
simply not sign themselves up at all and if they get extra time to
complete a task than it is just a bonus.

Do you think our mix of paid/volunteers is radically different from
other open-source projects?  There are plenty of projects that provide
estimates on release dates.  Many projects give these estimates far in
advance of the release and still deliver on the date, or adjust their
roadmap when things change.

> Or for those
> employed to work on svn, how much time they will have available? And
> has each of these employees specified their available time, what
> they're working on, and how long it will take them? Has all of this
> been collected across all of these volunteers into a workable plan?
> ... Nope.

Ok, you are in a room together.  Assemble the plan.  Why wouldn't
everyone in that room want to have an idea on how the remaining
problems are going to be tackled and possibly who is going to sign up
to take them?  Do you want to work on an issue only to see someone
else commit the same change before you do?  There are so few issues
left, that just by figuring out who plans to do the remaining items we
ought to be able to forecast a release date from that.

> Only if we knew everybody's time commitments. I don't know mine; I
> have no hard schedule. Do you have the schedules for the Collab.Net
> employees? That would solve part of the problem.

Yes.  Bert, Mike and Paul are all fully engaged on Subversion.  I
would say 40 hours a week but I think that does a disservice to the
amount of time they put into the project.  At some point, I imagine
Bert will turn attention to SharpSVN and AnkhSVN, but right now he
seems fully committed to wrapping up 1.7 issues.  Mike and Paul are
full time on Subversion.  If there are people that would have more
time to devote to 1.7 if they were being paid, feel free to contact me
privately.  I have already reached out to everyone I thought might
have some time though.

I could also ask Kamesh and his team to devote more time specifically
to 1.7 if there are specific tasks we want him to look at.


> Yes. People are not confident that we've capture everything in the
> issue tracker. Go ahead and apply "push" as you deem appropriate.
> We've been talking about it here, but we certainly haven't pushed each
> other so far.

By "push" I just mean that sometimes we hold a to-do list in our heads
or on our computers that we really should be pushing to the entire
community.  If anyone has items related to the release that they think
needs to be done, whether it be API review, docstring review, checking
our Apache release process etc..  We should create issues so that the
big picture of remaining work is in a single location we can all see.
That is all I mean by this.  I also think that by forcing people to
talk about this, maybe there will be a collective realization that we
are close to the release.


> Turn that around: what are YOU expecting. Hyrum said we would start
> cutting releases on June 1. What else would you like?

The pre-releases seem more for Hyrum to work out our Apache release
process than for anything else.  I want to know when to expect the
release candidate so that I can then have a rough approximation of the
GA.

> I think we're all aware of this. I also think that we're working on
> getting the last bits wrapped up. Maybe it isn't the form that you
> would like?

I appreciate the hard work people have been doing to finish the
release.  The issue close rate the last month has been very good.  I
want everyone to see that the finish line is right in front of us.  I
want you all to take advantage of this rare chance of being in the
same room to talk to each other and get consensus on the remaining
work.

> June 1, we'll start some early releases. So far, there isn't anything
> further to say, with confidence. If you have ideas on how to make
> statements with greater confidence, then please do so. I don't think
> it will be coming out of this room.

I think there are two primary things that you could accomplish by
being together and able to talk to each other:

1) Talk about what stands between where we are now and a finished
release.  Make note of any items not captured in the tracker and add
them.

2) Talk about some of these items in detail and the right approach to
solving them.

Doing #1 will give people confidence that there is not much left to
do, or it will at least reveal all the things we have not captured
already.  Doing #2 should give a sense on how much work it will take
and possibly who can do it.

-- 
Thanks

Mark Phippard
http://markphip.blogspot.com/

Re: First 1.7.0 prerelease: June 1

[Greg Stein <gs...@gmail.com>]

On Wed, May 18, 2011 at 07:43, Mark Phippard <ma...@gmail.com> wrote:
> On Wed, May 18, 2011 at 7:25 AM, Greg Stein <gs...@gmail.com> wrote:
>
>> If solving these were easy, then I believe they would have been done
>> by now. We're down to the hard issues, and people don't have a high
>> degree of confidence in planning when they would be done.
>
> We are not the only software project in the world working on hard
> problems or that finds it difficult to provide estimates.  It sucks,
> but part of being a professional is doing the best you can.

Do you think that you can plan around *volunteers*? Or for those
employed to work on svn, how much time they will have available? And
has each of these employees specified their available time, what
they're working on, and how long it will take them? Has all of this
been collected across all of these volunteers into a workable plan?
... Nope.

My point was that I don't see anybody doing this. If you'd like to
start collecting such a plan and hitting up each person to assemble
it, then please... by all means. I just don't see it happening from
the people here doing the coding.

>  We have
> single-digit issues.  We ought to be able to provide a date on when we
> think the release will be ready and then push each other to meet that
> deadline.

Only if we knew everybody's time commitments. I don't know mine; I
have no hard schedule. Do you have the schedules for the Collab.Net
employees? That would solve part of the problem.

> I also have a feeling that people think there are other items not
> captured in the tracker.  That may be true, and I am saying you need
> to push each other to unlock that information and get it captured in
> the tracker.

Yes. People are not confident that we've capture everything in the
issue tracker. Go ahead and apply "push" as you deem appropriate.
We've been talking about it here, but we certainly haven't pushed each
other so far.

>> I'm also not seeing anybody here standing up and whipping us into
>> coming up with a date. People seem pretty content with "we're working
>> on it as best we can."
>
> Consider yourself whipped :)

Doesn't work on me :-)

> Seriously though, who or what are you expecting?  A lot of people have

Turn that around: what are YOU expecting. Hyrum said we would start
cutting releases on June 1. What else would you like?

> probably just given up or are numb from the 2+ year release cycle we
> are on.  It is not just about declaring a date, hitting it and
> releasing the software.  It is also acknowledging there is an
> ecosystem downstream that uses the software and has to make their own
> schedules around it.

I think we're all aware of this. I also think that we're working on
getting the last bits wrapped up. Maybe it isn't the form that you
would like?

>  I would like to be able to do a Subclipse
> release concurrent with 1.7.  I need to be able to budget time to make
> the adjustments needed for 1.7 and handle my own release process.  We
> use Subversion in our enterprise products and need to be able to plan
> releases around making Subversion 1.7 available in those products,
>
> Give the community a date.  If it doesn't look like you can hit the
> date, then update the community with the new date and move on.  It
> seems like the shame in possibly missing a date is being given too
> much weight.  The community still wants and needs those dates.

June 1, we'll start some early releases. So far, there isn't anything
further to say, with confidence. If you have ideas on how to make
statements with greater confidence, then please do so. I don't think
it will be coming out of this room.

Cheers,
-g

Re: First 1.7.0 prerelease: June 1

[Mark Phippard <ma...@gmail.com>]

On Wed, May 18, 2011 at 7:25 AM, Greg Stein <gs...@gmail.com> wrote:

> If solving these were easy, then I believe they would have been done
> by now. We're down to the hard issues, and people don't have a high
> degree of confidence in planning when they would be done.

We are not the only software project in the world working on hard
problems or that finds it difficult to provide estimates.  It sucks,
but part of being a professional is doing the best you can.  We have
single-digit issues.  We ought to be able to provide a date on when we
think the release will be ready and then push each other to meet that
deadline.

I also have a feeling that people think there are other items not
captured in the tracker.  That may be true, and I am saying you need
to push each other to unlock that information and get it captured in
the tracker.

> I'm also not seeing anybody here standing up and whipping us into
> coming up with a date. People seem pretty content with "we're working
> on it as best we can."

Consider yourself whipped :)

Seriously though, who or what are you expecting?  A lot of people have
probably just given up or are numb from the 2+ year release cycle we
are on.  It is not just about declaring a date, hitting it and
releasing the software.  It is also acknowledging there is an
ecosystem downstream that uses the software and has to make their own
schedules around it.  I would like to be able to do a Subclipse
release concurrent with 1.7.  I need to be able to budget time to make
the adjustments needed for 1.7 and handle my own release process.  We
use Subversion in our enterprise products and need to be able to plan
releases around making Subversion 1.7 available in those products,

Give the community a date.  If it doesn't look like you can hit the
date, then update the community with the new date and move on.  It
seems like the shame in possibly missing a date is being given too
much weight.  The community still wants and needs those dates.

-- 
Thanks

Mark Phippard
http://markphip.blogspot.com/

Re: First 1.7.0 prerelease: June 1

[Greg Stein <gs...@gmail.com>]

On Wed, May 18, 2011 at 07:16, Mark Phippard <ma...@gmail.com> wrote:
> On Wed, May 18, 2011 at 6:17 AM, Hyrum K Wright <hy...@hyrumwright.org> wrote:
>> Fire up your testing environments: they're going to get some use over
>> the next couple of months.
>>
>> In an effort to release early and often, I'd like to cut the first
>> 1.7.0 pre-release on June 1.  This will likely be a beta release,
>> since there are still blocking issues, but I hope to get this into the
>> hands of users (and third-party consumers) soon-ish.
>>
>> From that point, we'll continue to release beta's until we branch,
>> which will be when the "1.7.0" issues are all closed.  After the
>> branch, we'll start the release candidates and the soak period.
>
> I am +1 on doing pre-release build(s) if you want to, but if this is
> the group statement coming out of Berlin this is pretty weak.  You
> have more SVN contributors in the room then there are open issues.  It
> seems like we ought to be able to make a statement with a good degree
> of confidence on when we will branch for the release and issue the
> release candidate.  If nothing else, we should be seeing a flood of
> new issues being added to the milestone.  If the people in the room
> cannot identify the reasons to not do the release .. and put them in
> the issue tracker.  Then what is blocking us?  Why can't the remaining
> issues all be closed by June 1?
>
> I realize the way you worded this statement leaves that open as a
> possibility, but with single digit open issues we ought to be able to
> make a stronger statement as to when we *think* we will be ready for
> release candidate.

If solving these were easy, then I believe they would have been done
by now. We're down to the hard issues, and people don't have a high
degree of confidence in planning when they would be done.

I'm also not seeing anybody here standing up and whipping us into
coming up with a date. People seem pretty content with "we're working
on it as best we can."

Cheers,
-g

Re: First 1.7.0 prerelease: June 1

[Mark Phippard <ma...@gmail.com>]

On Wed, May 18, 2011 at 6:17 AM, Hyrum K Wright <hy...@hyrumwright.org> wrote:
> Fire up your testing environments: they're going to get some use over
> the next couple of months.
>
> In an effort to release early and often, I'd like to cut the first
> 1.7.0 pre-release on June 1.  This will likely be a beta release,
> since there are still blocking issues, but I hope to get this into the
> hands of users (and third-party consumers) soon-ish.
>
> From that point, we'll continue to release beta's until we branch,
> which will be when the "1.7.0" issues are all closed.  After the
> branch, we'll start the release candidates and the soak period.

I am +1 on doing pre-release build(s) if you want to, but if this is
the group statement coming out of Berlin this is pretty weak.  You
have more SVN contributors in the room then there are open issues.  It
seems like we ought to be able to make a statement with a good degree
of confidence on when we will branch for the release and issue the
release candidate.  If nothing else, we should be seeing a flood of
new issues being added to the milestone.  If the people in the room
cannot identify the reasons to not do the release .. and put them in
the issue tracker.  Then what is blocking us?  Why can't the remaining
issues all be closed by June 1?

I realize the way you worded this statement leaves that open as a
possibility, but with single digit open issues we ought to be able to
make a stronger statement as to when we *think* we will be ready for
release candidate.

-- 
Thanks

Mark Phippard
http://markphip.blogspot.com/