Relationship between 23.1.90 pretest and 23.2 pretest?

Relationship between 23.1.90 pretest and 23.2 pretest?

[Karl Fogel]

We've just released a pretest.  In some places, it is referred to as the
"23.2 pretest" and in other places as "the 23.1.90" pretest.

These are the same thing right?  If not, what's their relationship?  (I
looked for, but couldn't find, a document about our release processes.)

[1] uses both names, apparently to refer to the same thing.  [2] is the
pretest announcement, which uses "23.1.90" consistently but also says
"This is the first pretest for what will be the Emacs 23.2 release",
which implies that "23.1.90 pretest" is the same as "23.2 pretest".

And the reason I'm asking, of course, is to check that we're ready to
finally do the bzr switchover.  The end of [2] implies that it is now
time to do that, but I wanted to make sure before starting the thread.

Thanks,
-Karl

[1] http://lists.gnu.org/archive/html/emacs-devel/2009-12/msg00186.html
[2] http://lists.gnu.org/archive/html/emacs-devel/2009-12/msg00390.html

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Chong Yidong]

> We've just released a pretest.  In some places, it is referred to as the
> "23.2 pretest" and in other places as "the 23.1.90" pretest.
>
> These are the same thing right?  If not, what's their relationship?  (I
> looked for, but couldn't find, a document about our release processes.)

The pretests leading up to Emacs 23.2 are 23.1.90, 23.1.91, 23.1.92,
etc.

> And the reason I'm asking, of course, is to check that we're ready to
> finally do the bzr switchover.  The end of [2] implies that it is now
> time to do that, but I wanted to make sure before starting the thread.

Yes, I am in favor of moving forward on the bzr switchover.

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Stefan Monnier]

>> We've just released a pretest.  In some places, it is referred to as the
>> "23.2 pretest" and in other places as "the 23.1.90" pretest.
>> 
>> These are the same thing right?  If not, what's their relationship?  (I
>> looked for, but couldn't find, a document about our release processes.)

> The pretests leading up to Emacs 23.2 are 23.1.90, 23.1.91, 23.1.92,
> etc.

Yes, in my mind "the pretest" is not a product but a process.  So the
"pretest for Emacs-23.2" includes releasing a bunch of versions, which
we traditionally name 23.1.9x (which can grow to 23.1.99x or to 23.1.10x
depending on who's in charge at the time we move past 23.1.99).
If I see "the 23.2 pretest code", that would usually mean "whichever
version was last distributed during the pretest".

A pretest version such as 23.1.93 becomes immediately obsolete the
moment 23.1.94 is cut.  Actually, not just obsolete: any mention (let
alone use) of 23.1.93 after 23.1.94 has been released is of very bad
taste and will be duly noted by Santa.

>> And the reason I'm asking, of course, is to check that we're ready to
>> finally do the bzr switchover.  The end of [2] implies that it is now
>> time to do that, but I wanted to make sure before starting the thread.
> Yes, I am in favor of moving forward on the bzr switchover.

Yes, now is a good time for it.  The only problems I can think of are:
- I'm leaving tomorrow for the country of Borges (and Maradona), so
  until January I will be less available.
- loggerhead doesn't seem to be working any more on Savannah.
- the commit-diff email system is not yet working, and I just learned
  that Savannah has something implemented for it, so it would make a lot
  of sense to try and use that.

Personally I'm not convinced those should stop us from switching now
since I expect the Xmas season to be a good "quiet" time for the switch:
1- declare a 100% freeze on any commits.
2- ask Andreas to perform one last conversion.
3- ask the Savannah admins to make the CVS repository read-only.
4- when the conversion is done and looks good: declare victory.
5- start rebuilding the country: get Loggerhead working, get
   commit-diffs emails working, get bzr+ssh working, ...


        Stefan

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Juanma Barranquero]

On Thu, Dec 10, 2009 at 18:05, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> - I'm leaving tomorrow for the country of Borges (and Maradona)

And Julio Cortázar, possibly the best short stories writer ever.

> 2- ask Andreas to perform one last conversion.

Why is that necessary? The current repository is maintained
up-to-date, isn't it?

    Juanma

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Eli Zaretskii]

> From: Juanma Barranquero <lekktu@gmail.com>
> Date: Fri, 11 Dec 2009 00:39:13 +0100
> Cc: Karl Fogel <kfogel@red-bean.com>, Chong Yidong <cyd@stupidchicken.com>,
> 	emacs-devel@gnu.org
> 
> On Thu, Dec 10, 2009 at 18:05, Stefan Monnier <monnier@iro.umontreal.ca> wrote:
> > 2- ask Andreas to perform one last conversion.
> 
> Why is that necessary? The current repository is maintained
> up-to-date, isn't it?

It is maintained up-to-date by Andreas, who does the synchronization,
conversion and whatever else is needed, manually once a day.

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Chong Yidong]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:

> Personally I'm not convinced those should stop us from switching now
> since I expect the Xmas season to be a good "quiet" time for the switch:
> 1- declare a 100% freeze on any commits.
> 2- ask Andreas to perform one last conversion.
> 3- ask the Savannah admins to make the CVS repository read-only.
> 4- when the conversion is done and looks good: declare victory.
> 5- start rebuilding the country: get Loggerhead working, get
>    commit-diffs emails working, get bzr+ssh working, ...

A good way to do this would be to just jump to step 3.  Then, whichever
time the Savannah admins get around to making the CVS repository
read-only, will be the time everyone needs to migrate.  Less
co-ordination involved

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Stefan Monnier]

> A good way to do this would be to just jump to step 3.  Then, whichever
> time the Savannah admins get around to making the CVS repository
> read-only, will be the time everyone needs to migrate.  Less
> co-ordination involved

OK, let's do that.


        Stefan "Waiting at Miami airport"

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Karl Fogel]

Stefan Monnier <monnier@IRO.UMontreal.CA> writes:
>> A good way to do this would be to just jump to step 3.  Then, whichever
>> time the Savannah admins get around to making the CVS repository
>> read-only, will be the time everyone needs to migrate.  Less
>> co-ordination involved
>
> OK, let's do that.

Agreed!

Shall I file an SR now, asking the admins to make the CVS repository
read-only?  I don't see one already -- but I don't want to step on
anyone's toes if one of you have already put in the request.

I just went through http://www.emacswiki.org/emacs/BzrForEmacsDevs to
test the recipe from absolute scratch.  Even over a wireless modem,
"bzr branch sftp://kfogel@bzr.savannah.gnu.org/srv/bzr/emacs/trunk"
took only 52 minutes.  For a one-time transition, that's quite
acceptable IMHO.

I fixed a few out-of-date things in the page.  I think we're good to go
now.  Once the admins have made the CVS repository read-only, we can ask
Andreas for one final conversion.  If there are any bugs remaining in
the workflow, we'll fix them in production.

Yidong is totally right that actually switching over will be what forces
us to get Loggerhead working, get commit-diff emails, bzr+ssh, etc :-).

-Karl "Not waiting at the Miami airport"

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Chong Yidong]

Karl Fogel <kfogel@red-bean.com> writes:

> Shall I file an SR now, asking the admins to make the CVS repository
> read-only?  I don't see one already -- but I don't want to step on
> anyone's toes if one of you have already put in the request.

Please do.

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Karl Fogel]

Chong Yidong <cyd@stupidchicken.com> writes:
> Karl Fogel <kfogel@red-bean.com> writes:
>> Shall I file an SR now, asking the admins to make the CVS repository
>> read-only?  I don't see one already -- but I don't want to step on
>> anyone's toes if one of you have already put in the request.
>
> Please do.

https://savannah.gnu.org/support/index.php?107170

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Jan Djärv]

I haven't read this thread. Did we just move to Bzr?  I missed the  
announcement.


      Jan D.


12 dec 2009 kl. 22.09 skrev Karl Fogel <kfogel@red-bean.com>:

> Chong Yidong <cyd@stupidchicken.com> writes:
>> Karl Fogel <kfogel@red-bean.com> writes:
>>> Shall I file an SR now, asking the admins to make the CVS repository
>>> read-only?  I don't see one already -- but I don't want to step on
>>> anyone's toes if one of you have already put in the request.
>>
>> Please do.
>
> https://savannah.gnu.org/support/index.php?107170
>

Re: Relationship between 23.1.90 pretest and 23.2 pretest?

[Karl Fogel]

Jan Djärv <jan.h.d@swipnet.se> writes:
> I haven't read this thread. Did we just move to Bzr?  I missed the
> announcement.

We are moving now.  

The first step is to make the CVS repository read-only.  Then Andreas
can do a clean final conversion, without worrying about, say, a
non-atomic multi-directory CVS commit being only partially included in
the conversion.

Thus there will be a brief period, while Andreas converts, when we can't
make commits to CVS but Bzr is not ready either.  It shouldn't be long;
I don't know exactly how long these conversions take, but my guess is a
day or less.  You can continue to commit to CVS until the repository
goes read-only, of course.

(I suppose there's an outside chance of a race condition, due to CVS's
non-atomicity, whereby some of a given commit gets in but not all,
because the admins make the repository read-only during the commit.  In
practice, I'm not worried about it.  There are ways they could avoid
even that small risk... but it's better just to let them do whatever is
easiest for them, so that it gets done sooner.)

-Karl

> 12 dec 2009 kl. 22.09 skrev Karl Fogel <kfogel@red-bean.com>:
>> Chong Yidong <cyd@stupidchicken.com> writes:
>>> Karl Fogel <kfogel@red-bean.com> writes:
>>>> Shall I file an SR now, asking the admins to make the CVS repository
>>>> read-only?  I don't see one already -- but I don't want to step on
>>>> anyone's toes if one of you have already put in the request.
>>>
>>> Please do.
>>
>> https://savannah.gnu.org/support/index.php?107170

Bazaar switchover: waiting for admins, how to speed it?

[Karl Fogel]

Karl Fogel <kfogel@red-bean.com> writes:
> Chong Yidong <cyd@stupidchicken.com> writes:
>> Karl Fogel <kfogel@red-bean.com> writes:
>>> Shall I file an SR now, asking the admins to make the CVS repository
>>> read-only?  I don't see one already -- but I don't want to step on
>>> anyone's toes if one of you have already put in the request.
>>
>> Please do.
>
> https://savannah.gnu.org/support/index.php?107170

Yidong or Stefan, do you know of anything we should do to get attention
on this ticket?  (There's no sign yet that any admins have seen it.)

Thanks,
-Karl