Forum: Ruby-core Schedule for the 1.8.7 release

Hi, developers,

After seeing Ruby 1.9.0-1 released earlier today, I, as the release
manager of the 1.8 series, would like to announce here that it is time
to start working on the next release, 1.8.7.

Below is the timeline I made up.  The reason for the slow start (long
period for discussion) is because I'm going on a week's trip next week
to Portland and Seattle.

Schedule for 1.8.7: (subject to change as it proceeds)

    Now - 2008-03-22
      Discuss what features should and/or can be backported, and get
      the release manager to make decisions as to which features are
      going to be worked on for 1.8.7.

    2008-03-23 18:00:00 JST
      Declare feature freeze.  The tracker will be keeping a complete
      list of bugs to fix and new features to implement.

    2008-03-24 - 2008-04-03
      Work on Tracker issues.  New features must be implemented by the
      end of this period, while some bugs may remain.

    2008-04-04 12:00:00 JST
      Declare code freeze.

      - The ruby_1_8_7 branch is forked off from the ruby_1_8 branch.

      - The ruby_1_8_7 branch is frozen from the beginning and any
        change on the branch requires an explicitly approval by the
        release manager.

      - The ruby_1_8 branch is not frozen, however, massive changes
        are forbidden for the moment.

    2008-04-05 18:00:00 JST
      Release 1.8.7-preview1

    2008-04-12 18:00:00 JST
      Release 1.8.7-preview2

    2008-04-20 18:00:00 JST
      Release 1.8.7.
      Lift code freeze.


I realize we have a lot of things to do because the difference between
1.9 and 1.8 (1.8.6) has become huge and so many changes have been made
to 1.8 since 1.8.6 was released.  That is why it is going to take this
long time to make 1.8.7 a decent release for mission critical scenes
yet an exciting and beloved release for ambitious users.

Please make sure to Cc me important mails, especially ones that need
my approval.  I tend to drop a ball on high-volume mailing lists
because of my poor fast reading skills.

For issue tracking, I made a tracker for 1.8.7 release engineering:

   http://rubyforge.org/tracker/?func=browse&group_id...

Any changes to the tracker issues will be forwarded to me, so you
don't have to bother pointing them to me via email.

Now, let's start the work!

Regards,

　ということで、 1.8.7 リリースに向けた作業を開始します。

　アナウンスに書いた通り、 1.8.6 からだいぶ間があいたため、これ
までに行われた変更を含めてテストしないといけないこと、 1.9 の
開発がかなり進行し、バックポートや互換性のための変更をこれから
行わねばならないことから、十分に長い準備期間を取る必要があると
判断しました。

　スケジュールに従って、まずは今月22日までにやることのリストを
まとめます。私は3/4〜10の間はおそらくろくに反応できませんが、
それ以外はこまめにチェックします。重要なメールは遠慮なくCcして
ください。

　やることがまとまったら、個々のタスクは Tracker で管理します。

　それでは、よろしくお願いします。

At Sat, 1 Mar 2008 20:58:00 +0900,

On Sat, Mar 01, 2008 at 08:58:00PM +0900, Akinori MUSHA wrote:
>     Now - 2008-03-22
>       Discuss what features should and/or can be backported, and get
>       the release manager to make decisions as to which features are
>       going to be worked on for 1.8.7.

Would it be possible to add the patch from [ruby-core:15662]?  I have a
lot of code that this change would greatly simplify.

I never saw a response the first time, otherwise I wouldn't ask twice.

Paul

Hi,

At Mon, 3 Mar 2008 15:01:53 -0500,
Paul Brannan wrote:
> On Sat, Mar 01, 2008 at 08:58:00PM +0900, Akinori MUSHA wrote:
> >     Now - 2008-03-22
> >       Discuss what features should and/or can be backported, and get
> >       the release manager to make decisions as to which features are
> >       going to be worked on for 1.8.7.
>
> Would it be possible to add the patch from [ruby-core:15662]?  I have a
> lot of code that this change would greatly simplify.

You need to give a convincing example or two out of "a lot" to show
how it reasonably helps you.  What's wrong with a static variable
instead?

> >       going to be worked on for 1.8.7.
> 
> Would it be possible to add the patch from [ruby-core:15662]? 
>  I have a lot of code that this change would greatly simplify.
> 
> I never saw a response the first time, otherwise I wouldn't ask twice.

Hi Paul,

Do you think that could eventually be used for method annotations in a
more general way? Is that what you ultimately have in mind? Or am I way
off here?

Regards,

Dan


This communication is the property of Qwest and may contain confidential 
or
privileged information. Unauthorized use of this communication is 
strictly
prohibited and may be unlawful.  If you have received this communication
in error, please immediately notify the sender by reply e-mail and 
destroy
all copies of the communication and any attachments.

On Tue, Mar 04, 2008 at 06:41:59AM +0900, Akinori MUSHA wrote:
> You need to give a convincing example or two out of "a lot" to show
> how it reasonably helps you.  What's wrong with a static variable
> instead?

1. In Rice (http://rice.rubyforge.org) we attach data to a function so
   that when the function is called we can know how to convert the
   arguments from Ruby types to C++ types.

2. In ruby-libjit (http://rubyforge.org/projects/ruby-libjit) I attach a
   jit-compiled function object to a method so that the function won't
   be destroyed unless the method it is implementing is removed.

3. In the ruby interpreter, struct.c could be extended to have O(1)
   access for all members of the struct, not just the first 10 (I have
   an unreleased extension that does just this).

In neither case is static data appropriate: in the first case, because
functions are being generated at compile-time, and in the second case,
because functions are being generated at run-time.

This was discussed in 2004 in old RCR#202, and Matz was concerned that
the solution was not re-entrant, but I don't believe this to be the
case.

Paul

On Tue, Mar 04, 2008 at 07:22:56AM +0900, Berger, Daniel wrote:
> Do you think that could eventually be used for method annotations in a
> more general way? Is that what you ultimately have in mind? Or am I way
> off here?

This particular implementation could not be used for general method
annotatations, because it only works for CFUNCs.  Implementing the
feature for other method types requires a much bigger change to the
interpreter.

However, I can imagine many uses for a generalized annotation mechanism.

Paul

Still packing..

At Mon, 3 Mar 2008 09:41:23 -0500,
Rocky Bernstein wrote:
> > Below is the timeline I made up.  The reason for the slow start (long
>
> In what forum are changes and features going to be discussed? ruby-core? As

Yes, ruby-core would be the best place to discuss in.

> for issues,
> see  http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/...
>
> Alas no comments/discussion. If you want me to work up a patch, let me know.
>
> As for back-ports, I have the honor of submitting the first tracker request
> :-)
>
> http://rubyforge.org/tracker/index.php?func=detail...

OK, I've read the thread and the patch looks fair.  But, is what you
mean by "a patch" different than is attached to the request?

I'm a little confused with what you are asking, so let me try to 
summarize.

I have two requests for the release so far. The first change request is 
a
patch by Mauricio Fernandez. It has been applied to Ruby 1.9 although
Mauricio also gave a patch for 1.8 and it is just as valid there.

The second change request stems from an observation that in conditionals
like "if", "while", or "until", a trace hook gets called twice for no 
good
reason that I can tell. Originally I had thought there was a stop before 
the
expression was evaluated and another after the expression was evaluated, 
but
that is not the case.

Here is an example:

x=1
x=2 if x=3

If you debug this you will see that the values of x are 1 and 2 even 
though
there is a stage in theory where x=3.

I'm not sure what the intended behavior is supposed to be, but if it is 
to
run the trace hook once, I think I could provide a patch for that.

On Ruby 1.9 if you run the above program, I do not believe you do not 
get
the duplicate call to the trace hook. Since Ruby 1.9 changes constantly, 
I
could be wrong; so folks please double check. I have been using a 
patched
debug.rb on 1.9 since the one I have installed on Ubuntu ruby
1.9.0(2007-08-30 patchlevel 0) [i486-linux] doesn't work. (No doubt
this been
subsequently fixed.)

> -----Original Message-----
> request is a patch by Mauricio Fernandez. It has been applied 
> to Ruby 1.9 although Mauricio also gave a patch for 1.8 and 
> it is just as valid there.

What patch was that?

Thanks,

Dan


This communication is the property of Qwest and may contain confidential 
or
privileged information. Unauthorized use of this communication is 
strictly
prohibited and may be unlawful.  If you have received this communication
in error, please immediately notify the sender by reply e-mail and 
destroy
all copies of the communication and any attachments.

On Mon, Mar 3, 2008 at 5:24 PM, Berger, Daniel <Daniel.Berger@qwest.com>
wrote:

> > I have two requests for the release so far. The first change
> > request is a patch by Mauricio Fernandez. It has been applied
> > to Ruby 1.9 although Mauricio also gave a patch for 1.8 and
> > it is just as valid there.
>
> What patch was that?


SCRIPT_LINES__ should not append when loading a file more than
once<%20http://rubyforge.org/tracker/index.php?func=detail...

Hi,

We have two more days before the feature (list) freeze.

    http://rubyforge.org/tracker/?atid=22040&group_id=...

Feel free to add more.  Note that, however, since we don't have much
time to review new code, you shouldn't expect that a complex feature
can go in, which is not yet implemented at all.

What I think we should concentrate on is bug/feature/documentation
fixes, backports and forward compatibility.

Regards,

At Sat, 1 Mar 2008 20:58:00 +0900,

　あまり時間が取れずぎりぎりになっての再告知ですみませんが、
1.8.7 に向けてのやることリストを今週いっぱいで締め切ります。

    http://rubyforge.org/tracker/?atid=22040&group_id=...

　時間上の制約から、まだ実装が存在しないような新機能を1.8に追加
するということは基本的にはしません。実装や機能の不備の修正、1.9
からのバックポート、そして1.9との前方互換性に関するものに集中
したいと思います。

　すでにいくつか上がって(自分でも上げて)いますが、ほかにも
要望があれば追加してください。

　よろしくお願いします。

It might be helpful if you went over the tracker list and make note of 
which
things are deferred and which are not. Specifically I'm interested in 
#18722
(trace hook getting called after expression evaluation ) also mentioned 
in
http://www.ruby-forum.com/topic/145818

Thanks.

$B1J0f!wCNG=!%6e9)Bg$G$9!%(B

From: "Akinori MUSHA" <knu@iDaemons.org>
Subject: [ruby-dev:34103] Re: Schedule for the 1.8.7 release
Date: Fri, 21 Mar 2008 23:25:26 +0900
Message-ID: <861w64f929.knu@iDaemons.org>
> $B!!$"$^$j;~4V$,<h$l$:$.$j$.$j$K$J$C$F$N:F9pCN$G$9$_$^$;$s$,!"(B
> 1.8.7 $B$K8~$1$F$N$d$k$3$H%j%9%H$r:#=5$$$C$Q$$$GDy$a@Z$j$^$9!#(B
> 
>     http://rubyforge.org/tracker/?atid=22040&group_id=...
> 
> $B!!;~4V>e$N@)Ls$+$i!"$^$@<BAu$,B8:_$7$J$$$h$&$J?75!G=$r(B1.8$B$KDI2C(B
> $B$9$k$H$$$&$3$H$O4pK\E*$K$O$7$^$;$s!#<BAu$d5!G=$NITHw$N=$@5!"(B1.9
> $B$+$i$N%P%C%/%]!<%H!"$=$7$F(B1.9$B$H$NA0J}8_49@-$K4X$9$k$b$N$K=8Cf(B
> $B$7$?$$$H;W$$$^$9!#(B

Tcl/Tk 8.5 $BBP1~$J$I$r4^$a$?(B Ruby/Tk 
$B$NBgI}2~=$$r5^$$$G$$$?$N$G$9$,!$(B
$B;DG0$J$,$i4V$K9g$$$^$;$s$G$7$?!%(B(;_;)
$B;EJ}$,$J$$$N$G(B 
($B=$@5$NItJ,E*$J@Z=P$7$O$A$g$C$HFq$7$$$N$G(B)$B!$(B
widget demo $B$N%P%0=$@5$@$1$r(B ruby_1_8 $B$K(B commit 
$B$7$^$7$?!%(B
$B$h$m$7$/$*4j$$CW$7$^$9!%(B

At Sat, 22 Mar 2008 01:43:50 +0900,
Rocky Bernstein wrote:
> It might be helpful if you went over the tracker list and make note of which
> things are deferred and which are not. Specifically I'm interested in #18722
> (trace hook getting called after expression evaluation ) also mentioned in
> http://www.ruby-forum.com/topic/145818

It has been handled, thanks.

By the way, bug fixes are not included in the feature requests here.
I am not going to set a dead-line on bug-fix requests any time soon.
Feel free to point out any time, do not hesitate to take advantage of
the release engineering period for getting bugs fixed.  That is
exactly what it is for! :)

Regards,

At Fri, 21 Mar 2008 23:16:54 +0900,
I wrote:
> We have two more days before the feature (list) freeze.
>
>     http://rubyforge.org/tracker/?atid=22040&group_id=...

We will be working on backporting the features that are currently on
the above list.

The deadline for the implementation to be done is the below:

    2008-04-04 12:00:00 JST
      Declare code freeze.

Patches and commits are always welcome, when we currently have only a
few committers working on backports.  If you have the code, you can
just follow up an issue with a patch attached.

Besides those new features, feel free to point me to bug reports that
are not handled yet.

Thanks for your help!

Akinori MUSHA wrote:
 > Besides those new features, feel free to point me to bug reports that
 > are not handled yet.

  Well, there is [ruby-core:15833] Module#dup give something strange.


Guy Decoux

At Mon, 24 Mar 2008 21:39:45 +0900,
I wrote:
> The deadline for the implementation to be done is the below:
>
>     2008-04-04 12:00:00 JST
>       Declare code freeze.

As I have not been able to spend time for Ruby the last couple of
weeks due to my personal problems to cope with, the deadline will be
put off until next Thursday.

Sorry for the inconvenience, and thanks for your support!

On Mon, Mar 24, 2008 at 2:39 PM, Akinori MUSHA <knu@idaemons.org> wrote:
>  Besides those new features, feel free to point me to bug reports that
>  are not handled yet.

I'd like to have this small issue fixed:
http://rubyforge.org/tracker/index.php?func=detail...

Basically this prevents reusing Net::POP objects - once it obtains
mail count from the server,
it never checks it again, not even in another session.

Thanks.
Jano

Index: pop.rb
===================================================================
--- pop.rb  (revision 15689)
+++ pop.rb  (working copy)
@@ -577,6 +577,8 @@

     def do_finish
       @mails = nil
+      @n_mails = nil
+      @n_bytes = nil
       @command.quit if @command
     ensure
       @started = false

Hi,

In message "Re: Schedule for the 1.8.7 release"
    on Fri, 4 Apr 2008 04:31:40 +0900, "Jano Svitok" 
<jan.svitok@gmail.com> writes:

|I'd like to have this small issue fixed:
|http://rubyforge.org/tracker/index.php?func=detail...
|
|Basically this prevents reusing Net::POP objects - once it obtains
|mail count from the server,
|it never checks it again, not even in another session.

Done.  Thank you.

              matz.

Hi,

Sorry for the delay again, but it will require some more time for me
to get all backports done.

The first preview of 1.8.7 will hopefully be released during this
weekend, about one week late behind the initially proposed schedule.

Regards,

At Thu, 3 Apr 2008 18:04:19 +0900,

Are all the methods (Object.tap, Proc.curry, Symbol.to_proc and so on)
listed at

   http://rubyforge.org/tracker/?atid=22040&group_id=...

still planned for inclusion in 1.8.7?  If so, I'll make a note of it in
the reprint for my book.

Thanks,

  David Flanagan

Sorry for replying too late,

At Fri, 11 Apr 2008 08:40:10 +0900,
David Flanagan wrote:
> Are all the methods (Object.tap, Proc.curry, Symbol.to_proc and so on)
> listed at
> 
>   http://rubyforge.org/tracker/?atid=22040&group_id=...
> 
> still planned for inclusion in 1.8.7?  If so, I'll make a note of it
> in the reprint for my book.

The ones currently in NEWS and some Enumerable/Enumerator/Array
methods and String#{start,end}_with? will be in time for 1.8.7.  For
other things I'm not sure.

Actually I had to spend time this weekend for miscellaneous stuff and
I've just started working on Ruby.  The first preview is getting
delayed again but will be available in 12 hours.  I'll be working hard
tonight!

Regards,

Hi,

I have finally done most of the backporting tasks.  Yes, finally.

I have to admit that I totally underestimated the amount of tasks as
it turned out there were a lot more things to do than I expected.

I still have some leftovers to tackle, so I yet again put off the
preview1 until tomorrow evening.  I will stop massive backports
shortly and concentrate on bugfixing and nitpicking, then roll a
preview release cooperating with core developers.

Thank you for your patience!