[ruby-core:22803] Relegate 1.8.6 to Engine Yard, part II

[ruby-core:22803] Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Hello and sorry for my being slow for this issue.  It's OK now for me to pass
stewardship to Engine Yard.  I've backported what's necessary.

To all:

Can you test latest 1.8.6 repository?  I want to do one last release of 1.8.6
before I give up, and I'd be happy if that happens in this month.  I believe I
didn't break anything but I know I'm far from being perfect.  So please, check
if it works.

To Ezra and Engine Yard people:

Are you ready?  I think you need to take at least
(1) svn account on svn.ruby-lang.org to commit
(2) ssh account on ftp.ruby-lang.org to put files there, and
(3) web account on redmine.ruby-lang.org to control issues.

[ruby-core:22816] Re: Relegate 1.8.6 to Engine Yard, part II

[Ryan Davis]

On Mar 10, 2009, at 00:57 , Urabe Shyouhei wrote:

> Hello and sorry for my being slow for this issue.  It's OK now for  
> me to pass
> stewardship to Engine Yard.  I've backported what's necessary.
>
> To all:
>
> Can you test latest 1.8.6 repository?  I want to do one last release  
> of 1.8.6
> before I give up, and I'd be happy if that happens in this month.  I  
> believe I
> didn't break anything but I know I'm far from being perfect.  So  
> please, check
> if it works.

YAY! thanks!

Would it be possible to get that file_load_ok patch backported before  
you release?

[ruby-core:22817] Re: Relegate 1.8.6 to Engine Yard, part II

[Roger Pack]

> Can you test latest 1.8.6 repository?  I want to do one last release of 1.8.6
> before I give up, and I'd be happy if that happens in this month.  I believe I
> didn't break anything but I know I'm far from being perfect.  So please, check
> if it works.

Has this bug been addressed?
http://redmine.ruby-lang.org/issues/show/1223
Thanks!
-=r

[ruby-core:22831] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Urabe Shyouhei wrote:
> Can you test latest 1.8.6 repository?

Here's a rubyspec output (of 9fcd252b92f5b2130db17a924e7338f2a9773ec1
at master) on my machine:

http://gist.github.com/77259

1) this spec is broken.
2) this was a bug of ruby. cf [ruby-dev:37019]
3) this spec is broken.
4) this spec is broken.
5) this spec is broken.
6) this spec is broken.
7) this was a bug of ruby. cf [ruby-core:19632]

[ruby-core:22832] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Roger Pack wrote:
> Has this bug been addressed?
> http://redmine.ruby-lang.org/issues/show/1223
> Thanks!

Seems not (in the whole project).

[ruby-core:22833] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Ryan Davis wrote:
> 
> On Mar 10, 2009, at 00:57 , Urabe Shyouhei wrote:
> 
>> Hello and sorry for my being slow for this issue.  It's OK now for me
>> to pass
>> stewardship to Engine Yard.  I've backported what's necessary.
>>
>> To all:
>>
>> Can you test latest 1.8.6 repository?  I want to do one last release
>> of 1.8.6
>> before I give up, and I'd be happy if that happens in this month.  I
>> believe I
>> didn't break anything but I know I'm far from being perfect.  So
>> please, check
>> if it works.
> 
> YAY! thanks!
> 
> Would it be possible to get that file_load_ok patch backported before
> you release?

Yes. I'll work on it today.

[ruby-core:22840] Re: Relegate 1.8.6 to Engine Yard, part II

[brian ford]

Hi,

On Tue, Mar 10, 2009 at 6:52 PM, Urabe Shyouhei <shyouhei@ruby-lang.org> wrote:
> Urabe Shyouhei wrote:
>> Can you test latest 1.8.6 repository?
>
> Here's a rubyspec output (of 9fcd252b92f5b2130db17a924e7338f2a9773ec1
> at master) on my machine:
>
> http://gist.github.com/77259
>
> 1) this spec is broken.
> 2) this was a bug of ruby. cf [ruby-dev:37019]
> 3) this spec is broken.
> 4) this spec is broken.
> 5) this spec is broken.
> 6) this spec is broken.
> 7) this was a bug of ruby. cf [ruby-core:19632]
>
>

Thanks, we'll fix the specs. And thanks for the reference to the bug reports.

Brian

[ruby-core:22842] Re: Relegate 1.8.6 to Engine Yard, part II

[Roger Pack]

Looks like it was just fixed.  Thank you nobu


> Has this bug been addressed?
> http://redmine.ruby-lang.org/issues/show/1223

[ruby-core:22845] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Urabe Shyouhei wrote:
>> Would it be possible to get that file_load_ok patch backported before
>> you release?
> 
> Yes. I'll work on it today.

Ryan, I hav'nt been able to reproduce this bug on 1.8.6.  I suspect it doesn't
hurt 1.8.7 and earlier.  Can you check?

zsh % mkdir -p a/foo b

zsh % touch b/foo
zsh % ls -ld a/* b/*
drwxr-xr-x 2 shyouhei shyouhei 40 2009-03-11 21:13 a/foo
-rw-r--r-- 1 shyouhei shyouhei  0 2009-03-11 21:13 b/foo
zsh % ruby -v -Ia -Ib -e 'load "foo"'
ruby 1.8.6 (2009-03-10 patchlevel 362) [x86_64-linux]
a/foo: Is a directory - a/foo (Errno::EISDIR)
        from -e:1:in `load'
        from -e:1

[ruby-core:22847] Re: Relegate 1.8.6 to Engine Yard, part II

[Ezra Zygmuntowicz]

Hi~

On Mar 10, 2009, at 12:57 AM, Urabe Shyouhei wrote:

> Hello and sorry for my being slow for this issue.  It's OK now for  
> me to pass
> stewardship to Engine Yard.  I've backported what's necessary.
>
> To all:
>
> Can you test latest 1.8.6 repository?  I want to do one last release  
> of 1.8.6
> before I give up, and I'd be happy if that happens in this month.  I  
> believe I
> didn't break anything but I know I'm far from being perfect.  So  
> please, check
> if it works.
>
> To Ezra and Engine Yard people:
>
> Are you ready?  I think you need to take at least

Yes we are ready, we are in the process of hiring a full time person  
to handle this role, so I think we will be ready pretty much anytime  
in the next week or two after you roll another 1.8.6 release.


> (1) svn account on svn.ruby-lang.org to commit
> (2) ssh account on ftp.ruby-lang.org to put files there, and
> (3) web account on redmine.ruby-lang.org to control issues.

Yup, please let me know how to proceed on these issues, feel free to  
contact me offlist with details. We look forward to making sure flks  
who need it have a stable 1.8.6 to use for the forseeable future

Thanks

Ezra Zygmuntowicz
ez@engineyard.com

[ruby-core:22848] Re: Relegate 1.8.6 to Engine Yard, part II

[daz]

Ezra Zygmuntowicz wrote:
> Hi~
>
> On Mar 10, 2009, at 12:57 AM, Urabe Shyouhei wrote:
>
>> (1) svn account on svn.ruby-lang.org to commit
>> (2) ssh account on ftp.ruby-lang.org to put files there, and
>> (3) web account on redmine.ruby-lang.org to control issues.
>
>
> Thanks
>
> Ezra Zygmuntowicz
> ez@engineyard.com
>
>

Hi Ezra,

A thought you might want to discuss with Shyouhei:

I'm assuming that, in future, svn.ruby-lang.org remains
the official repository for 1.8.6 and that it will have a
single authorised committer == the dcommit command of:
http://www.kernel.org/pub/software/scm/git/docs/git-svn.html
via your EY git repo?


Let's say that Matz discovers a bug in a library in 1.8.7 ...

   raise 'Security Alert' unless intrusions = 0   # silly example

which also exists in 1.8.6 branch.

The easiest thing for him to do may be to commit the fix to
1.8.6 svn at the same time as 1.8.7 to help you.
If I'm understanding, git-svn (fetch) will handle that as well
but would it be the right protocol? Or should write access to
1.8.6 svn be closed to all except 'dcommit'? And would Matz
(or any other svn committer) need to raise a ticket against
1.8.6 on Redmine, instead?

I'm not asking for a reply because it's none of my business
but it seems like an issue that's likely to arise some time.

Fascinating. Good luck to all concerned. :)


daz

[ruby-core:22857] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Ezra Zygmuntowicz wrote:
>> To Ezra and Engine Yard people:
>>
>> Are you ready?  I think you need to take at least
> 
> Yes we are ready, we are in the process of hiring a full time person to
> handle this role, so I think we will be ready pretty much anytime in the
> next week or two after you roll another 1.8.6 release.
> 
> 
>> (1) svn account on svn.ruby-lang.org to commit
>> (2) ssh account on ftp.ruby-lang.org to put files there, and
>> (3) web account on redmine.ruby-lang.org to control issues.
> 
> Yup, please let me know how to proceed on these issues, feel free to
> contact me offlist with details. We look forward to making sure flks who
> need it have a stable 1.8.6 to use for the forseeable future

Great, then take a look at:
http://redmine.ruby-lang.org/wiki/ruby/CommitterHowto

[ruby-core:22907] Re: Relegate 1.8.6 to Engine Yard, part II

[Nobuyoshi Nakada]

Hi,

At Wed, 11 Mar 2009 21:20:41 +0900,
Urabe Shyouhei wrote in [ruby-core:22845]:
> Ryan, I hav'nt been able to reproduce this bug on 1.8.6.  I suspect it doesn't
> hurt 1.8.7 and earlier.  Can you check?
> 
> zsh % mkdir -p a/foo b
> 
> zsh % touch b/foo
> zsh % ls -ld a/* b/*
> drwxr-xr-x 2 shyouhei shyouhei 40 2009-03-11 21:13 a/foo
> -rw-r--r-- 1 shyouhei shyouhei  0 2009-03-11 21:13 b/foo
> zsh % ruby -v -Ia -Ib -e 'load "foo"'
> ruby 1.8.6 (2009-03-10 patchlevel 362) [x86_64-linux]
> a/foo: Is a directory - a/foo (Errno::EISDIR)
>         from -e:1:in `load'
>         from -e:1

It is exactly the reported behavior.  Now 1.8.8dev ignores
a/foo because it is a directory, but loads b/foo instead.

To be backported or not is another story.  Honestly, I can't
imagine why merb has to create such directory structure.

-- 
Nobu Nakada

[ruby-core:22910] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Nobuyoshi Nakada wrote:
> It is exactly the reported behavior.  Now 1.8.8dev ignores
> a/foo because it is a directory, but loads b/foo instead.

OK, I see the point now.  Let me take a closer look at the code...

[ruby-core:22917] Re: Relegate 1.8.6 to Engine Yard, part II

[Charles Oliver Nutter]

Nobuyoshi Nakada wrote:
> At Wed, 11 Mar 2009 21:20:41 +0900,
> Urabe Shyouhei wrote in [ruby-core:22845]:
>> zsh % ruby -v -Ia -Ib -e 'load "foo"'
>> ruby 1.8.6 (2009-03-10 patchlevel 362) [x86_64-linux]
>> a/foo: Is a directory - a/foo (Errno::EISDIR)
>>         from -e:1:in `load'
>>         from -e:1
> 
> It is exactly the reported behavior.  Now 1.8.8dev ignores
> a/foo because it is a directory, but loads b/foo instead.

Has anything been added to rubyspec for the new behavior?

- Charlie

[ruby-core:22919] Re: Relegate 1.8.6 to Engine Yard, part II

[Ryan Davis]

On Mar 15, 2009, at 22:43 , Nobuyoshi Nakada wrote:

> It is exactly the reported behavior.  Now 1.8.8dev ignores
> a/foo because it is a directory, but loads b/foo instead.
>
> To be backported or not is another story.  Honestly, I can't
> imagine why merb has to create such directory structure.

I can't speak for merb, but the issue I'm seeing is:

ZenTest-4.0.0/bin/autotest

vs:

cucumber-0.1.16/lib/autotest/discover.rb

rubygems installs "bin" scripts with a gem wrapper:

> require 'rubygems'
>
> version = ">= 0"
>
> if ARGV.first =~ /^_(.*)_$/ and Gem::Version.correct? $1 then
>   version = $1
>   ARGV.shift
> end
>
> gem 'ZenTest', version
> load 'autotest'

Unfortunately, that directory structure is here to stay.

[ruby-core:23021] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Urabe Shyouhei wrote:
> Nobuyoshi Nakada wrote:
>> It is exactly the reported behavior.  Now 1.8.8dev ignores
>> a/foo because it is a directory, but loads b/foo instead.
> 
> OK, I see the point now.  Let me take a closer look at the code...

I think I've backported what was necessary.

[ruby-core:23023] Re: Relegate 1.8.6 to Engine Yard, part II

[Luis Lavena]

On Thu, Mar 26, 2009 at 4:45 AM, Urabe Shyouhei <shyouhei@ruby-lang.org> wrote:
> Urabe Shyouhei wrote:
>> Nobuyoshi Nakada wrote:
>>> It is exactly the reported behavior.  Now 1.8.8dev ignores
>>> a/foo because it is a directory, but loads b/foo instead.
>>
>> OK, I see the point now.  Let me take a closer look at the code...
>
> I think I've backported what was necessary.
>

What about backporting the mkmf changes in #1127

http://redmine.ruby-lang.org/issues/show/1127

Those are needed by One-Click Installer. I cannot build any new 1.8.6
version because of that.

Thank you.
-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

[ruby-core:23031] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Luis Lavena wrote:
> What about backporting the mkmf changes in #1127

Done.  But can you check to see if I've made mistakes?  I believe not, but I
don't have a working MinGW environment.

[ruby-core:23073] Re: Relegate 1.8.6 to Engine Yard, part II

[Luis Lavena]

On Fri, Mar 27, 2009 at 8:07 AM, Urabe Shyouhei <shyouhei@ruby-lang.org> wrote:
> Luis Lavena wrote:
>> What about backporting the mkmf changes in #1127
>
> Done.  But can you check to see if I've made mistakes?  I believe not, but I
> don't have a working MinGW environment.
>

Hello over there.

It seems #1127 is not working as expected.

After looking at lib/mkmf.rb, it seems that one small piece of code is
interfering.

See the inlined patch to fix it:

Index: lib/mkmf.rb
===================================================================
--- lib/mkmf.rb	(revision 23100)
+++ lib/mkmf.rb	(working copy)
@@ -1063,7 +1063,7 @@
       if CONFIG['target_os'] != 'cygwin'
         vpath.each {|p| p.sub!(/.*/, '$(shell cygpath -u \&)')}
       end
-    when 'msdosdjgpp', 'mingw32'
+    when 'msdosdjgpp'
       CONFIG['PATH_SEPARATOR'] = ';'
     end
   end

After applied that patch, I can successfully go beyond Win32API
extension compilation. yay!

Luis@KEORE (D:\Users\Luis\projects\oss\oci\rubyinstaller)
$ sandbox\ruby_build\ruby.exe -v
ruby 1.8.6 (2009-03-27 patchlevel 367) [i386-mingw32]

That change is needed for me to update One-Click Installer builds.

Please consider it before release 1.8.6.

Thank you,
-- 
Luis Lavena
AREA 17
-
Perfection in design is achieved not when there is nothing more to add,
but rather when there is nothing more to take away.
Antoine de Saint-Exupéry

[ruby-core:23080] Re: Relegate 1.8.6 to Engine Yard, part II

[Urabe Shyouhei]

Luis Lavena wrote:
> Please consider it before release 1.8.6.

Sorry for a late reply, actually this was the last thing I added to the
release.  I think you can build from the tarballs.