From jano at dimaion.com  Fri Dec  2 14:39:49 2011
From: jano at dimaion.com (Damian Janowski)
Date: Fri, 2 Dec 2011 16:39:49 -0300
Subject: Weird random Zbatery exception
Message-ID: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>

Hello everyone,

I'm giving Zbatery a try, currently testing with ThreadPool.

By the way, this instance is sitting behind Varnish.

I can't figure out precise steps to trigger this exception, but it
happens after a few GET/POST requests:

/usr/local/lib/ruby/gems/1.9.1/gems/unicorn-4.1.1/lib/unicorn/http_server.rb:361:in
`awaken_master': undefined method `kgio_trywrite' for nil:NilClass
(NoMethodError)
        from /usr/local/lib/ruby/gems/1.9.1/gems/unicorn-4.1.1/lib/unicorn/http_server.rb:130:in
`block in start'
        from /usr/local/lib/ruby/gems/1.9.1/gems/rainbows-4.3.1/lib/rainbows/thread_pool.rb:32:in
`call'
        from /usr/local/lib/ruby/gems/1.9.1/gems/rainbows-4.3.1/lib/rainbows/thread_pool.rb:32:in
`join'
        from /usr/local/lib/ruby/gems/1.9.1/gems/rainbows-4.3.1/lib/rainbows/thread_pool.rb:32:in
`block in worker_loop'
        from /usr/local/lib/ruby/gems/1.9.1/gems/rainbows-4.3.1/lib/rainbows/thread_pool.rb:30:in
`each'
        from /usr/local/lib/ruby/gems/1.9.1/gems/rainbows-4.3.1/lib/rainbows/thread_pool.rb:30:in
`worker_loop'
        from /usr/local/lib/ruby/gems/1.9.1/gems/zbatery-4.1.1/lib/zbatery.rb:100:in
`join'
        from /usr/local/lib/ruby/gems/1.9.1/gems/zbatery-4.1.1/bin/zbatery:121:in
`<top (required)>'
        from /usr/local/bin/zbatery:19:in `load'
        from /usr/local/bin/zbatery:19:in `<main>'

(Needless to say, I googled already without much luck).

Any ideas?

D.

From normalperson at yhbt.net  Fri Dec  2 17:47:11 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Fri, 2 Dec 2011 14:47:11 -0800
Subject: Weird random Zbatery exception
In-Reply-To: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
Message-ID: <20111202224711.GA23422@dcvr.yhbt.net>

Damian Janowski <jano at dimaion.com> wrote:
> I'm giving Zbatery a try, currently testing with ThreadPool.

Thanks for giving it a try! :)

> I can't figure out precise steps to trigger this exception, but it
> happens after a few GET/POST requests:

> /usr/local/lib/ruby/gems/1.9.1/gems/unicorn-4.1.1/lib/unicorn/http_server.rb:361:in
> `awaken_master': undefined method `kgio_trywrite' for nil:NilClass
> (NoMethodError)

Does your application fork/exec other processes?  (`` or "system()",
popen all included).

> (Needless to say, I googled already without much luck).

I think you found a new bug, does the following patch fix it?

diff --git a/lib/zbatery.rb b/lib/zbatery.rb
index 735b1e6..17d7c06 100644
--- a/lib/zbatery.rb
+++ b/lib/zbatery.rb
@@ -82,6 +82,7 @@ module Rainbows
       trap(:USR1) { Thread.new { reopen_logs } }
       trap(:USR2) { Thread.new { reexec } }
       trap(:HUP) { Thread.new { reexec; stop } }
+      trap(:CHLD, "IGNORE")
 
       # technically feasible in some cases, just not sanely supportable:
       %w(TTIN TTOU WINCH).each do |sig|

From normalperson at yhbt.net  Fri Dec  2 18:40:03 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Fri, 2 Dec 2011 15:40:03 -0800
Subject: Weird random Zbatery exception
In-Reply-To: <20111202224711.GA23422@dcvr.yhbt.net>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
	<20111202224711.GA23422@dcvr.yhbt.net>
Message-ID: <20111202234003.GA27418@dcvr.yhbt.net>

Eric Wong <normalperson at yhbt.net> wrote:
> diff --git a/lib/zbatery.rb b/lib/zbatery.rb
> index 735b1e6..17d7c06 100644
> --- a/lib/zbatery.rb
> +++ b/lib/zbatery.rb
> @@ -82,6 +82,7 @@ module Rainbows
>        trap(:USR1) { Thread.new { reopen_logs } }
>        trap(:USR2) { Thread.new { reexec } }
>        trap(:HUP) { Thread.new { reexec; stop } }
> +      trap(:CHLD, "IGNORE")

On second thought, "IGNORE" will probably break existing code.
trap(:CHLD, "DEFAULT") should be better.

Btw, this issue shouldn't affect Rainbows!, only Zbatery.
Can you confirm?

From normalperson at yhbt.net  Fri Dec  2 20:46:32 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Fri, 2 Dec 2011 17:46:32 -0800
Subject: Weird random Zbatery exception
In-Reply-To: <20111202234003.GA27418@dcvr.yhbt.net>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
	<20111202224711.GA23422@dcvr.yhbt.net>
	<20111202234003.GA27418@dcvr.yhbt.net>
Message-ID: <20111203014632.GA3639@dcvr.yhbt.net>

Eric Wong <normalperson at yhbt.net> wrote:
> Eric Wong <normalperson at yhbt.net> wrote:
> > +      trap(:CHLD, "IGNORE")
> 
> On second thought, "IGNORE" will probably break existing code.
> trap(:CHLD, "DEFAULT") should be better.

I just pushed out Zbatery 4.1.1.2.g3693 (prerelease) to RubyGems.org
with trap(:CHLD,"DEFAULT") + test cases.

You can install it with: gem install --pre zbatery

Let me know if it fixes your problem since I've been meaning to release
for 90d1d769b9af83d6d1 anyways (unlink oldbin pid file during graceful
shutdown).

(I also found https://redmine.ruby-lang.org/issues/5700 while working
 on this :o)

From jano at dimaion.com  Mon Dec  5 13:02:10 2011
From: jano at dimaion.com (Damian Janowski)
Date: Mon, 5 Dec 2011 15:02:10 -0300
Subject: Weird random Zbatery exception
In-Reply-To: <20111202224711.GA23422@dcvr.yhbt.net>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
	<20111202224711.GA23422@dcvr.yhbt.net>
Message-ID: <CABAijKFJ4wigHt68_na+yFz=Tgv+y8tC1kU=X0q+jwy9JwXULA@mail.gmail.com>

On Fri, Dec 2, 2011 at 7:47 PM, Eric Wong <normalperson at yhbt.net> wrote:
> Thanks for giving it a try! :)

Thank you, Eric, for your prompt reply!

> Does your application fork/exec other processes? ?(`` or "system()",
> popen all included).

Not in my application code, and I'm pretty sure not in any of the few
libraries I'm using.

From jano at dimaion.com  Mon Dec  5 13:02:44 2011
From: jano at dimaion.com (Damian Janowski)
Date: Mon, 5 Dec 2011 15:02:44 -0300
Subject: Weird random Zbatery exception
In-Reply-To: <20111202234003.GA27418@dcvr.yhbt.net>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
	<20111202224711.GA23422@dcvr.yhbt.net>
	<20111202234003.GA27418@dcvr.yhbt.net>
Message-ID: <CABAijKFZCQUYyjip6-AZpEgjddZr59Qb3UOhCFcpHmDiKDqB_A@mail.gmail.com>

On Fri, Dec 2, 2011 at 8:40 PM, Eric Wong <normalperson at yhbt.net> wrote:
> Btw, this issue shouldn't affect Rainbows!, only Zbatery.
> Can you confirm?

I can confirm Rainbows has been stable for a few days (whereas Zbatery
crashed after a few requests).

I'll test the pre release now and let you know, thank you again.

From jano at dimaion.com  Mon Dec  5 13:29:24 2011
From: jano at dimaion.com (Damian Janowski)
Date: Mon, 5 Dec 2011 15:29:24 -0300
Subject: Weird random Zbatery exception
In-Reply-To: <20111203014632.GA3639@dcvr.yhbt.net>
References: <CABAijKG5VwHN9p9p_4R4oSruFVH7j3TY0Z6twxLSLpKoxcO8jw@mail.gmail.com>
	<20111202224711.GA23422@dcvr.yhbt.net>
	<20111202234003.GA27418@dcvr.yhbt.net>
	<20111203014632.GA3639@dcvr.yhbt.net>
Message-ID: <CABAijKFzKjpKcVpUj3eu0Db+s5PwiOEe8S9HthYaQFY_HUtncA@mail.gmail.com>

On Fri, Dec 2, 2011 at 10:46 PM, Eric Wong <normalperson at yhbt.net> wrote:
> I just pushed out Zbatery 4.1.1.2.g3693 (prerelease) to RubyGems.org
> with trap(:CHLD,"DEFAULT") + test cases.
>
> You can install it with: gem install --pre zbatery
>
> Let me know if it fixes your problem since I've been meaning to release
> for 90d1d769b9af83d6d1 anyways (unlink oldbin pid file during graceful
> shutdown).

It does seem to fix the problem.

Thank you very much!

From normalperson at yhbt.net  Mon Dec  5 13:51:30 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Mon, 5 Dec 2011 18:51:30 +0000
Subject: [ANN] Zbatery 4.1.2 - we don't fork, but our apps may!
Message-ID: <20111205185130.GA23161@dcvr.yhbt.net>

* http://zbatery.bogomip.org/
* rainbows-talk at rubyforge.org
* git://bogomips.org/zbatery.git
* http://zbatery.bogomip.org/NEWS.atom.xml

Changes:

There are two bugfixes in this release.  Rack applications that
use fork() internally should now behave as-expected when
receiving SIGCHLD.  The pid file is also unlinked during a
graceful shutdown.

Thanks to Damian Janowski for the bug report and confirming
the fix!

From danny.tran at digisynd.com  Mon Dec  5 18:42:57 2011
From: danny.tran at digisynd.com (Danny Tran)
Date: Mon, 5 Dec 2011 17:42:57 -0600
Subject: Rainbows! + Nginx
Message-ID: <CAM4_PZdCVss61emzKwA_4Mg5x6vzeP9UY9OuhezTM3MYzdAuYw@mail.gmail.com>

I'm trying to run a Rails app that utilizes HTTP streaming with
Rainbows! behind Nginx.
Here is my nginx.conf:https://gist.github.com/a749361acd09b6ae80e2
Despite setting proxy_buffering off; it appears that nginx is still
buffering the output because I do not get a streaming response. ?When
I curl the app directly (port 3000) I get a proper streaming response,
but when I curl nginx (which proxies to the Rails app) I do not get a
streaming response.
Any thoughts?
Thanks,Danny

From normalperson at yhbt.net  Mon Dec  5 19:51:04 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Tue, 6 Dec 2011 00:51:04 +0000
Subject: Rainbows! + Nginx
In-Reply-To: <CAM4_PZdCVss61emzKwA_4Mg5x6vzeP9UY9OuhezTM3MYzdAuYw@mail.gmail.com>
References: <CAM4_PZdCVss61emzKwA_4Mg5x6vzeP9UY9OuhezTM3MYzdAuYw@mail.gmail.com>
Message-ID: <20111206005104.GA20671@dcvr.yhbt.net>

Danny Tran <danny.tran at digisynd.com> wrote:
> I'm trying to run a Rails app that utilizes HTTP streaming with
> Rainbows! behind Nginx.
> Here is my nginx.conf:https://gist.github.com/a749361acd09b6ae80e2
> Despite setting proxy_buffering off; it appears that nginx is still
> buffering the output because I do not get a streaming response. ?When
> I curl the app directly (port 3000) I get a proper streaming response,
> but when I curl nginx (which proxies to the Rails app) I do not get a
> streaming response.
> Any thoughts?

Does your Rack app actually set "X-Accel-Buffering: no" in a response
header?  Your config mentions it:

               # The Rack application may also set "X-Accel-Buffering (yes|no)"
               # in the response headers do disable/enable buffering on a
               # per-response basis.
               proxy_buffering off;
               proxy_pass_header X-Accel-Buffering;

I've mostly forgotten about Rails 3.1 streaming myself :x

I know current nginx versions only send HTTP/1.0 requests to the
backends, which means you need to force the response to stream
via "X-Accel-Buffering: no" since HTTP/1.0 doesn't have chunked
responses,

You should also try using "curl --http1.0" to better simulate nginx
behavior with HTTP/1.0 for testing.

From admail-3 at you.saikyou.biz  Mon Dec 12 03:26:59 2011
From: admail-3 at you.saikyou.biz (=?ISO-2022-JP?B?GyRCNWE/TT5wSnMkTiFKM3QhSyVTJTglTSU5JUglaSU5JUglNyU5JUYlYCU6GyhC?=)
Date: 12 Dec 2011 17:26:59 +0900
Subject: =?ISO-2022-JP?B?GyRCOl9CcCVvITwlKyE8NV5KZ0IuSnMbKEIhGyRCIVolRyE8JT9Gfk5PNkhMM0JnTkw8dUNtJE4kPyRhIVslYSVrJV4lLBsoQlNPSE8bJEJAODNoJUslZSE8JTkbKEI=?=
Message-ID: <20111212082659.8018.qmail@www9.bizmail.jp>

??????SOHO??????14,494??2011?12?? ??????:87,799
??????????????????????

???????????????????????????????????????????? 
???????? Please e-mail the delivery stop here.? kyo at biz-career.rulez.jp
?1?  ????SOHO????????????????????????????????
?PR????????????????????????????????
?????????????????! SOHO???????????????!
????????????????????!!???????????????????!
????????????????????????????????!?

??? !! ?????????? ? http://www.soho-comu.rash.jp/
?PR????????????????????????????????
???????12???????????????
 
??????????????????????????
??SOHO??????????3???????????????
??????????????????????????

?1????????????????! ????????????!
  ?????????????????????????????
 
?2?????????????????????????! ??????!
  ?SOHO????????????????!
 
?3??????????????????????!?????!
  ????????????????????!
      ?
      ? ?????????
      ? ???????! ??????????!! ?
      ?
      ?????  http://www.soho-comu.rash.jp/
 
????????????????
?????????????????
 
??????????????????????????
????SOHO??????!!
 
??????????????
 
1.???? ? 18,500 ? 5,550?
2.?????????????? ? 10,550 ? 45,900?
3.????????? ? 1,000 ? 31,531?
4.????????? 6 ? 42,000?
5.?????????? 4 ? 13,500?
6.???????????????? ? 90 ? 45,000?
 
????????183,481?????
???????????????????(1?2?4????)???????!!
 
     ?  http://www.soho-comu.rash.jp/
     ?    ?(??????????)
 
?????????????????????????
???????????????! ????????????!!
 
???????????????????????????????????
?????????????????????????SOHO??????????
??????????????????????????????????????????
???????????????????????????????????

?<???>??????????????
??????????????????????????????????????? 
??????????????????????

??????????????????????????????????
???????????????????????????

???????????????????????????????????????????? 
???????? Please e-mail the delivery stop here.? kyo at biz-career.rulez.jp

???????????????????????????????????
??????????????)??????????????Web??????????
???????????????????????????????????
Copyright (c) 1999  ???????????????????????All rights reserved.???????

From godfat at godfat.org  Thu Dec 29 19:28:25 2011
From: godfat at godfat.org (Lin Jen-Shin (godfat))
Date: Fri, 30 Dec 2011 08:28:25 +0800
Subject: Weird app error: uninitialized fiber (FiberError)
Message-ID: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>

Hello Rainbows,

I was trying to `use :CoolioFiberSpawn` with `Rainbows.sleep`
to simulate asynchrony HTTP requests. However, I keep seeing
weird "app error: uninitialized fiber (FiberError)".

I spent a few hours (ouch) tracking down to a simple case.
Suppose I have a Rack app which has `Rainbows.sleep`
inside a each block:
https://gist.github.com/1536857 (I am running Ruby 1.9.3)

class Weird
  def call env=nil
    [1].each{ # if you comment out the each block, then it works
      Rainbows.sleep 2
    }
    [200, {}, ["ok\n"]]
  end
end

Then `run Weird.new` would always cause

    app error: uninitialized fiber (FiberError)

But if I took out the each block
inside Weird#call, such like:

class Weird
  def call env=nil
    Rainbows.sleep 2
    [200, {}, ["ok\n"]]
  end
end

Then it works fine. Moreover, if I run the app like this:

    run lambda{ |env| Weird.new.call }

Then both with or without the each block would work.

Could you please explain this for me? Or is it a bug somewhere?
Thanks!!

From normalperson at yhbt.net  Thu Dec 29 19:44:13 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Thu, 29 Dec 2011 16:44:13 -0800
Subject: Weird app error: uninitialized fiber (FiberError)
In-Reply-To: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
References: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
Message-ID: <20111230004413.GA12870@dcvr.yhbt.net>

"Lin Jen-Shin (godfat)" <godfat at godfat.org> wrote:
> I was trying to `use :CoolioFiberSpawn` with `Rainbows.sleep`
> to simulate asynchrony HTTP requests. However, I keep seeing
> weird "app error: uninitialized fiber (FiberError)".

<snip>

> Could you please explain this for me? Or is it a bug somewhere?

This looks like a Ruby bug, possibly related to combining Fibers with
fork()...

Can you reproduce this with zbatery?  (zbatery is Rainbows!, without fork()),
just s/rainbows/zbatery/ in the command-line).

This may be related to https://bugs.ruby-lang.org/issues/5700

From godfat at gmail.com  Thu Dec 29 19:55:11 2011
From: godfat at gmail.com (Lin Jen-Shin (godfat))
Date: Fri, 30 Dec 2011 08:55:11 +0800
Subject: Weird app error: uninitialized fiber (FiberError)
In-Reply-To: <20111230004413.GA12870@dcvr.yhbt.net>
References: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
	<20111230004413.GA12870@dcvr.yhbt.net>
Message-ID: <CANhhkiGe7Eqw1CupBY0JSkzD1MMuuX6aN7hUGoMHK3F43fcbCA@mail.gmail.com>

On Fri, Dec 30, 2011 at 8:44 AM, Eric Wong <normalperson at yhbt.net> wrote:
> This looks like a Ruby bug, possibly related to combining Fibers with
> fork()...
>
> Can you reproduce this with zbatery? ?(zbatery is Rainbows!, without fork()),
> just s/rainbows/zbatery/ in the command-line).
>
> This may be related to https://bugs.ruby-lang.org/issues/5700

Thanks for the quick response.

I can reproduce this under zbatery as well, so I guess this is
not related to fork()? My ruby and gem versions:

ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin10.8.0]
zbatery (4.1.2)
rainbows (4.3.1)
unicorn (4.1.1)
cool.io (1.1.0)

I am not sure if I should be happy if this is really a ruby's bug.
I thought I just couldn't figure out how do fibers work, looking
through a lot of sources and doing a lot of experiments.

From normalperson at yhbt.net  Thu Dec 29 20:35:47 2011
From: normalperson at yhbt.net (Eric Wong)
Date: Fri, 30 Dec 2011 01:35:47 +0000
Subject: Weird app error: uninitialized fiber (FiberError)
In-Reply-To: <CANhhkiGe7Eqw1CupBY0JSkzD1MMuuX6aN7hUGoMHK3F43fcbCA@mail.gmail.com>
References: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
	<20111230004413.GA12870@dcvr.yhbt.net>
	<CANhhkiGe7Eqw1CupBY0JSkzD1MMuuX6aN7hUGoMHK3F43fcbCA@mail.gmail.com>
Message-ID: <20111230013547.GA30808@dcvr.yhbt.net>

"Lin Jen-Shin (godfat)" <godfat at gmail.com> wrote:
> I can reproduce this under zbatery as well, so I guess this is
> not related to fork()? My ruby and gem versions:
> 
> ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-darwin10.8.0]

It may be an issue with LLVM/GCC on your platform, I've seen
several issues on the ruby-core list related to that setup.

Maybe recompiling Ruby (and all extensions) with a non-LLVM
version of GCC will solve this.

> I am not sure if I should be happy if this is really a ruby's bug.
> I thought I just couldn't figure out how do fibers work, looking
> through a lot of sources and doing a lot of experiments.

Can you reproduce the issue on a GNU/Linux system?
I cannot support non-Free platforms, anyways.

From godfat at godfat.org  Thu Dec 29 21:20:19 2011
From: godfat at godfat.org (Lin Jen-Shin (godfat))
Date: Fri, 30 Dec 2011 10:20:19 +0800
Subject: Weird app error: uninitialized fiber (FiberError)
In-Reply-To: <20111230013547.GA30808@dcvr.yhbt.net>
References: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
	<20111230004413.GA12870@dcvr.yhbt.net>
	<CANhhkiGe7Eqw1CupBY0JSkzD1MMuuX6aN7hUGoMHK3F43fcbCA@mail.gmail.com>
	<20111230013547.GA30808@dcvr.yhbt.net>
Message-ID: <CAA2_N1t-EVYTEQjjbmhA4cX9oTbyGFQLn5zWc8vbra-iXmkmZA@mail.gmail.com>

On Fri, Dec 30, 2011 at 9:35 AM, Eric Wong <normalperson at yhbt.net> wrote:
> Maybe recompiling Ruby (and all extensions) with a non-LLVM
> version of GCC will solve this.

I'll try to recompile Ruby (and all extensions) tomorrow. (I really need to
sleep now) It seems now Apple does not provide a non-LLVM GCC now.
My /usr/bin/gcc is a link to llvm-gcc-4.2.

> Can you reproduce the issue on a GNU/Linux system?
> I cannot support non-Free platforms, anyways.

I should have thought this. Just tried on my Arch Linux:
ruby 1.9.3p0 (2011-10-30 revision 33570) [x86_64-linux]
I can't reproduce that on the machine.

That's a good news. We're deploying to Linux anyway.
Thanks!

(I am not sure how to narrow this down further to file a
bug on Ruby's issue tracker. Maybe I'll just show them
what I have now)

From godfat at godfat.org  Fri Dec 30 12:03:59 2011
From: godfat at godfat.org (Lin Jen-Shin (godfat))
Date: Sat, 31 Dec 2011 01:03:59 +0800
Subject: Weird app error: uninitialized fiber (FiberError)
In-Reply-To: <20111230013547.GA30808@dcvr.yhbt.net>
References: <CAA2_N1vcNCgKKr8RQ7QDmsudShHD6g5ONnAhzCxocDyBkTesVg@mail.gmail.com>
	<20111230004413.GA12870@dcvr.yhbt.net>
	<CANhhkiGe7Eqw1CupBY0JSkzD1MMuuX6aN7hUGoMHK3F43fcbCA@mail.gmail.com>
	<20111230013547.GA30808@dcvr.yhbt.net>
Message-ID: <CAA2_N1sktajtj37gHxTcgmH8BaScEmuQUfVugcUx_Fik_f-4GQ@mail.gmail.com>

On Fri, Dec 30, 2011 at 9:35 AM, Eric Wong <normalperson at yhbt.net> wrote:
> It may be an issue with LLVM/GCC on your platform, I've seen
> several issues on the ruby-core list related to that setup.
>
> Maybe recompiling Ruby (and all extensions) with a non-LLVM
> version of GCC will solve this.

Just want to let you know, after trying GCC 4.6 without LLVM
involved and reinstalling all binary extensions, I can no longer
reproduce this.

What's amazing is, when I switch back to the old Ruby which
is compiled with Apple's llvm-gcc, I can't reproduce that either!

O_o

Maybe it's restarting the computer would really solve everything,
or I just did something wrong somewhere... Then I guess I shouldn't
file a bug for this.

Thanks for your help!