[GRASS-dev] 6.4.4 planning

GRASS GRASS developer
21

On Thu, Jun 5, 2014 at 5:47 PM, Moritz Lennert <moritz.lennert@ulb.ac.be> wrote:
...

As I've already told Markus off-list, I think that in order for this to work
we would need a clearly defined release-process with clear announcements of
the different steps (pre-freeze warning sufficiently early, freeze
announcement, pre-tag announcement, etc) and a strict respect of the rules
during this process (i.e. no non bugfix commits during a freeze), etc.

It doesn't seem to happen.

Please note ... the OSGeoLive feature freeze for the FOSS4G Portland
conf is upcoming. Would be a pity to not even see 6.4.4 therein!

I'm not convinced that a time-based release-policy is the solution, but I
think a more clearly defined release-process policy might help in avoiding
frustrations.

So let's see then...

Markus

22

On 06/06/14 21:43, Markus Neteler wrote:

On Thu, Jun 5, 2014 at 5:47 PM, Moritz Lennert <moritz.lennert@ulb.ac.be> wrote:
...

As I've already told Markus off-list, I think that in order for this to work
we would need a clearly defined release-process with clear announcements of
the different steps (pre-freeze warning sufficiently early, freeze
announcement, pre-tag announcement, etc) and a strict respect of the rules
during this process (i.e. no non bugfix commits during a freeze), etc.

It doesn't seem to happen.

At this stage there have never been clear announcements, easily identifiable as such for all developers. We should not assume that every developer reads every thread. Thus, such announcements should be outside of existing threads, and clearly marked in the subject.

Please note ... the OSGeoLive feature freeze for the FOSS4G Portland
conf is upcoming. Would be a pity to not even see 6.4.4 therein!

I'm a bit more available in the next weeks, so I can help with this process. Just let me know.

Moritz

23

On Tue, Jun 10, 2014 at 3:01 PM, Moritz Lennert
<mlennert@club.worldonline.be> wrote:

On 06/06/14 21:43, Markus Neteler wrote:

On Thu, Jun 5, 2014 at 5:47 PM, Moritz Lennert <moritz.lennert@ulb.ac.be>
wrote:
...

As I've already told Markus off-list, I think that in order for this to
work
we would need a clearly defined release-process with clear announcements
of
the different steps (pre-freeze warning sufficiently early, freeze
announcement, pre-tag announcement, etc) and a strict respect of the
rules
during this process (i.e. no non bugfix commits during a freeze), etc.

It doesn't seem to happen.

[I referred to the lack of responses in a sarcastic way :slight_smile: ]

At this stage there have never been clear announcements, easily identifiable
as such for all developers. We should not assume that every developer reads
every thread. Thus, such announcements should be outside of existing
threads, and clearly marked in the subject.

Sure. But I may mention that I do these releases since end of 1997 and
any attempt in the past to implement a *respected* release schedule
failed! See the archive for older attempts.

Please note ... the OSGeoLive feature freeze for the FOSS4G Portland
conf is upcoming. Would be a pity to not even see 6.4.4 therein!

I'm a bit more available in the next weeks, so I can help with this process.
Just let me know.

Unless we don't start tagging one day it will never reach the OSGeoLive DVD.
Their freeze is the 6 July 2014 according to Cameron's email. And indeed we need
to have GRASS 7 in it for our upcoming workshop but that's offtopic
for this email thread.

Markus (who feels a bit blocked by some developers)

24

On Tue, Jun 10, 2014 at 3:30 PM, Markus Neteler <neteler@osgeo.org> wrote:

>> Please note ... the OSGeoLive feature freeze for the FOSS4G Portland
>> conf is upcoming. Would be a pity to not even see 6.4.4 therein!
>
> I'm a bit more available in the next weeks, so I can help with this
process.
> Just let me know.

Unless we don't start tagging one day it will never reach the OSGeoLive
DVD.
Their freeze is the 6 July 2014 according to Cameron's email. And indeed
we need
to have GRASS 7 in it for our upcoming workshop but that's offtopic
for this email thread.

It's off topic but that's actually a valid note. Do we aim on having 6.4 or
7.0 in next OSGeo Live. We cannot have both, in

[Workshops] [Live-demo] OSGeoLive in FOSS4G 2014 workshops?
http://lists.osgeo.org/pipermail/conference-workshops/2014-June/000538.html

Cameron Shorter says:

To date we only have one version of each application installed on

OSGeo-Live, and I don't think we should break that policy.

(We aim to ensure a clear and simple experience for new users).
So I think we should only have one version of GRASS installed.

So 6.4 or 7.0. And this might be complication for QGIS GRASS plugin.

And back to the main topic, I'm bit lost in 6.4 releasing, is something
blocking the release? (Except for limited time Markus have for it.)

Vaclav

25

On Tue, Jun 10, 2014 at 10:12 PM, Vaclav Petras <wenzeslaus@gmail.com> wrote:
...

And back to the main topic, I'm bit lost in 6.4 releasing, is something
blocking the release? (Except for limited time Markus have for it.)

I had enough time in the past two weeks.

The blocker is the wish of Moritz to start a full announcement cycle
of dealing with the release management which has been historically
ignored. Also his suggestion has been ignored so far, not giving much
hope for a real one.

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Markus

26

2014-06-10 23:07 GMT+02:00 Markus Neteler <neteler@osgeo.org>:

The blocker is the wish of Moritz to start a full announcement cycle
of dealing with the release management which has been historically
ignored. Also his suggestion has been ignored so far, not giving much
hope for a real one.

That's completely unrealistic. We can start such procedure with G7
(time-based release policy).

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other days/weeks/months.

Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

27

On Tue, Jun 10, 2014 at 5:15 PM, Martin Landa <landa.martin@gmail.com>
wrote:

2014-06-10 23:07 GMT+02:00 Markus Neteler <neteler@osgeo.org>:

> The blocker is the wish of Moritz to start a full announcement cycle
> of dealing with the release management which has been historically
> ignored. Also his suggestion has been ignored so far, not giving much
> hope for a real one.

That's completely unrealistic. We can start such procedure with G7
(time-based release policy).

> Personally I believe that getting-things-done is the better approach
> for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other
days/weeks/months.

For me email with "6.4.4*planning" was the announcement of the current

one. We for sure want some fixed time-based release cycle for 7 but there
is no reason to not release if it is ready just because of missing policy
(there will be next release soon, right? so you everybody can push his
changes there).

Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

28

Hi,

2014-06-10 23:29 GMT+02:00 Vaclav Petras <wenzeslaus@gmail.com>:

For me email with "6.4.4*planning" was the announcement of the current one.
We for sure want some fixed time-based release cycle for 7 but there is no
reason to not release if it is ready just because of missing policy (there
will be next release soon, right? so you everybody can push his changes
there).

I complete agree, lets tag 6.4.4RC1 today/tomorrow. Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

29

And back to the main topic, I'm bit lost in 6.4 releasing, is something

blocking the release?

as there are no blockers, the most simple thing is just tag and prepare RC1
let the nice work of coding into the wild.

-----
best regards
Helmut
--
View this message in context: http://osgeo-org.1560.x6.nabble.com/Re-6-4-4-planning-tp5094052p5145171.html
Sent from the Grass - Dev mailing list archive at Nabble.com.

30

On 10/06/14 23:07, Markus Neteler wrote:

On Tue, Jun 10, 2014 at 10:12 PM, Vaclav Petras <wenzeslaus@gmail.com> wrote:
...

And back to the main topic, I'm bit lost in 6.4 releasing, is something
blocking the release? (Except for limited time Markus have for it.)

I had enough time in the past two weeks.

The blocker is the wish of Moritz to start a full announcement cycle
of dealing with the release management which has been historically
ignored. Also his suggestion has been ignored so far, not giving much
hope for a real one.

I never intended for my suggestion to be a blocker. I had hoped it would help apease the conflicts that seemed to arise around the 6.4.4 release.

I still believe that such a more clearly defined process (and again, I'm _not_ speaking about fixed-date releases) would help, but apparently I'm the only one who thinks so.

Moritz

31

Hi,

2014-06-11 10:01 GMT+02:00 Moritz Lennert <mlennert@club.worldonline.be>:

I still believe that such a more clearly defined process (and again, I'm
_not_ speaking about fixed-date releases) would help, but apparently I'm the

could you please explain a bit, how do you want to set up clear
release process without defined timeline?

only one who thinks so.

You are not the only one, but it takes some time to discuss and set up
(ideally as RFC). So it's topic for GRASS 7 not GRASS 6, thats all.

Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

32

On 11/06/14 10:30, Martin Landa wrote:

Hi,

2014-06-11 10:01 GMT+02:00 Moritz Lennert <mlennert@club.worldonline.be>:

I still believe that such a more clearly defined process (and again, I'm
_not_ speaking about fixed-date releases) would help, but apparently I'm the

could you please explain a bit, how do you want to set up clear
release process without defined timeline?

I did not say without defined timeline within the process, but not fixed-date. IOW, if we want to keep up a minimum with the reputation of fairly stable software, I don't think that for a project such as GRASS it would be a good idea to say "we'll have a release every six months, one in January and one in July". Rather, I think that to avoid the discussions we've had concerning which types of code modifications should go into a specific branch at which times and who decides what goes in and what does not, we should define the procedure more clearly. This also includes a clear calendar of the process itself.

I'll draft an RFC.

Moritz

33

2014-06-11 11:29 GMT+02:00 Moritz Lennert <mlennert@club.worldonline.be>:

I did not say without defined timeline within the process, but not
fixed-date. IOW, if we want to keep up a minimum with the reputation of

ah, OK. Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

34

2014-06-10 23:15 GMT+02:00 Martin Landa <landa.martin@gmail.com>:

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other days/weeks/months.

I think we can go ahead... Martin

--
Martin Landa * http://geo.fsv.cvut.cz/gwiki/Landa

35

On Wed, Jun 11, 2014 at 9:02 PM, Martin Landa <landa.martin@gmail.com> wrote:

2014-06-10 23:15 GMT+02:00 Martin Landa <landa.martin@gmail.com>:

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other days/weeks/months.

I think we can go ahead... Martin

I'll wait for Moritz' ok since he objected to let me tag it.

Markus (offline on Thursday)

36

On 12/06/14 00:15, Markus Neteler wrote:

On Wed, Jun 11, 2014 at 9:02 PM, Martin Landa <landa.martin@gmail.com> wrote:

2014-06-10 23:15 GMT+02:00 Martin Landa <landa.martin@gmail.com>:

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other days/weeks/months.

I think we can go ahead... Martin

I'll wait for Moritz' ok since he objected to let me tag it.

I don't know where you got that idea from. I proposed a more structured process of doing it, that's all. Since this proposal has not met with any concrete feedback, other than it shouldn't be applied for 6.4, I'll just get out of the way and let you handle things the way you want to.

Let me just add (gratuitously, I'm aware of that, but I'm a bit tired of a process where people make clear and concrete suggestions on an issue, but the content of these suggestions is just ignored, and the people who made them are then blamed that nothing advances) that if you had applied the proposed procedure when it was proposed [1], we would be really close to RC2 by now...

Moritz

[1] http://lists.osgeo.org/pipermail/grass-dev/2014-May/069126.html

37

On Thu, Jun 12, 2014 at 9:20 AM, Moritz Lennert
<mlennert@club.worldonline.be> wrote:

On 12/06/14 00:15, Markus Neteler wrote:

On Wed, Jun 11, 2014 at 9:02 PM, Martin Landa <landa.martin@gmail.com>
wrote:

2014-06-10 23:15 GMT+02:00 Martin Landa <landa.martin@gmail.com>:

Personally I believe that getting-things-done is the better approach
for 6.4.4 RC1.

Let's tag RC1, otherwise we stay on the same place for other
days/weeks/months.

I think we can go ahead... Martin

I'll wait for Moritz' ok since he objected to let me tag it.

I don't know where you got that idea from.

I got that from email from you:

-------
On Mon, Jun 2, 2014 at 10:41 AM, Moritz Lennert
<mlennert@club.worldonline.be> wrote:

On 31/05/14 23:19, Markus Neteler wrote:

On Mon, May 26, 2014 at 12:58 PM, Moritz Lennert

...

- RC1 beginning of next week

Should that happen then?

I think (also in line with the above) a clear freeze announcement has to go
out first to give a 24-48h warning to everyone. Then RC1 can be tagged.

Sorry, but I won't be of much help these days as we've started exams here
and I'll be busy with that almost all week.

-------

... it exceeded as I explained to you my available time window.
Perhaps some people underestimate the amount of work to prepare a
release, even a RC. So I just remind you that a bit more than just
launching the "svn tag" is involved. And the number of broken releases
in the past is close to zero.

I proposed a more structured process of doing it, that's all.

... which is generally fine.

Since this proposal has not met with any
concrete feedback, other than it shouldn't be applied for 6.4, I'll just get
out of the way and let you handle things the way you want to.

We have 19 "virtual" full time developers according to ohloh. I cannot
speak for the others but I want to get RC1 done since it is *only* a
release candidate and nothing else. Then we can consider to even apply
a new procedure. Just please let us do this RC1 - the planning
accouncement is from mid of Nov 2013, so probably enough time ago.
Surely I can retired from GRASS 6 maintenance if that opens up things.
But let me do RC1 first :slight_smile:

Let me just add (gratuitously, I'm aware of that, but I'm a bit tired of a
process where people make clear and concrete suggestions on an issue, but
the content of these suggestions is just ignored, and the people who made
them are then blamed that nothing advances) that if you had applied the
proposed procedure when it was proposed [1], we would be really close to RC2
by now...

Yes, if more feedback would come I would be happy as well.

Markus

Moritz

[1] http://lists.osgeo.org/pipermail/grass-dev/2014-May/069126.html

38

Hi Moritz, all,

On Thu, Jun 12, 2014 at 10:20 PM, Markus Neteler <neteler@osgeo.org> wrote:

On Thu, Jun 12, 2014 at 9:20 AM, Moritz Lennert

...

if you had applied the proposed procedure when it was proposed [1],
we would be really close to RC2 by now...

If "you" means me: as pointed out I was travelling, now travelling
again. I had time before when responses were slow.

BTW: I forgot to mention that release management had a long been existing:

http://grasswiki.osgeo.org/wiki/Category:Release_Roadmap
- GRASS 6.1 Feature Plan
- GRASS 6.2 Feature Plan
- GRASS 6.3 Feature Plan
- GRASS 6.4 Feature Plan
--> http://trac.osgeo.org/grass/wiki/Grass6Planning

As you see, nobody cared in the past 2 years. That's how it is. Or was?
Of course I am happy to make a new attempt, maybe it works out nowadays.

Best
Markus

39

On 12/06/14 22:20, Markus Neteler wrote:

On Thu, Jun 12, 2014 at 9:20 AM, Moritz Lennert
<mlennert@club.worldonline.be> wrote:

On 12/06/14 00:15, Markus Neteler wrote:

[...]

I'll wait for Moritz' ok since he objected to let me tag it.

I don't know where you got that idea from.

I got that from email from you:

-------
On Mon, Jun 2, 2014 at 10:41 AM, Moritz Lennert

I think (also in line with the above) a clear freeze announcement has to go
out first to give a 24-48h warning to everyone. Then RC1 can be tagged.

This says "I think" so is just a voicing of opinion, it's not a veto. If voicing a suggestion is seen as blocking progress, then that's a sad state of affairs.

But let me do RC1 first :slight_smile:

I'm not standing in your way, especially in light of the limited time we all seem to have.

Moritz

40

On Fri, Jun 13, 2014 at 1:10 PM, Moritz Lennert
<mlennert@club.worldonline.be> wrote:
...

I'm not standing in your way, especially in light of the limited time we all
seem to have.

I'll try to do prepare RC1 later today or tomorrow.

Markus