From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.3.2 (2011-06-06) on dcvr.yhbt.net
X-Spam-Level: 
X-Spam-ASN: AS47066 71.19.144.0/20
X-Spam-Status: No, score=-1.2 required=3.0 tests=BAYES_00,FREEMAIL_FROM,
 FROM_STARTS_WITH_NUMS,MSGID_FROM_MTA_HEADER shortcircuit=no autolearn=no
 version=3.3.2
Path: news.gmane.org!not-for-mail
From: Hleb Valoshka <375gnu@gmail.com>
Newsgroups: gmane.comp.lang.ruby.kgio.general
Subject: Re: Concurrency issue in TestKgioUnixConnect?
Date: Thu, 5 Sep 2013 23:53:03 +0300
Message-ID:
 <CAAB-Kck2ED1-hbfKvkug2R+H+HhyBAvQhZq6qF3LRCeVwi9dnQ@mail.gmail.com>
References: <20130901140301.GI1272@loar>
 <CAAB-KckfeHu1JWpfA3b-RRycr4x5b=dxAKfBpUrxXTYYLt6tRg@mail.gmail.com>
 <20130901195835.GB30612@dcvr.yhbt.net>
 <CAAB-KckA9R_7ssHmDkoHqS1f1=FXqdY4giwN_jGEC5TxrvNz2Q@mail.gmail.com>
 <20130902212429.GA18310@dcvr.yhbt.net>
 <CAAB-Kcmd8nndbLhrjE12YtWfaJEbKh-_8w7G_E0qgQROBvT-=Q@mail.gmail.com>
NNTP-Posting-Host: plane.gmane.org
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
X-Trace: ger.gmane.org 1378414395 5922 80.91.229.3 (5 Sep 2013 20:53:15 GMT)
X-Complaints-To: usenet@ger.gmane.org
NNTP-Posting-Date: Thu, 5 Sep 2013 20:53:15 +0000 (UTC)
To: kgio@librelist.org
Original-X-From: kgio@librelist.org Thu Sep 05 22:53:19 2013
Return-path: <kgio@librelist.org>
Envelope-to: gclrkg-kgio@m.gmane.org
List-Archive: <http://librelist.com/archives/kgio/>
List-Help: <http://librelist.com/help.html>
List-Id: <kgio-public@bogomips.org>
List-Post: <mailto:kgio@librelist.org>
List-Subscribe: <mailto:kgio-subscribe@librelist.com>
List-Unsubscribe: <mailto:kgio-unsubscribe@librelist.com>
Precedence: list
Original-Sender: kgio@librelist.org
Xref: news.gmane.org gmane.comp.lang.ruby.kgio.general:213
Archived-At:
 <http://permalink.gmane.org/gmane.comp.lang.ruby.kgio.general/213>
Received: from zedshaw2.xen.prgmr.com ([71.19.156.177]) by plane.gmane.org
 with esmtp (Exim 4.69) (envelope-from <kgio@librelist.org>) id
 1VHgYI-0007qB-18 for gclrkg-kgio@m.gmane.org; Thu, 05 Sep 2013 22:53:18 +0200
Received: from zedshaw2.xen.prgmr.com (unknown [IPv6:::1]) by
 zedshaw2.xen.prgmr.com (Postfix) with ESMTP id 79AA173E30 for
 <gclrkg-kgio@m.gmane.org>; Thu,  5 Sep 2013 21:02:22 +0000 (UTC)

On 9/3/13, Hleb Valoshka <375gnu@gmail.com> wrote:
> I was able to reproduce it only one or two times in virtualized
> GNU/kFreeBSD.

Problem is more interesting. I decided to check your latest changes
for test_poll.rb under kfreebsd-i386 and i'd got failure in
test_unix_connect.rb with teardown of test_unix_socket_new_invalid:

  1) Error:
test_unix_socket_new_invalid(TestKgioUnixConnect):
Errno::ENOENT: No such file or directory -
/tmp/kgio_unix_120130906-13330-m89438/kgio_unix_120130906-13330-ch4t73
    test/test_unix_connect.rb:33:in `unlink'
    test/test_unix_connect.rb:33:in `teardown'

The same failure as on ia64, mipsel and x32 Jeremy pointed in the
first message. And it definitely not a race condition because for the
next time i run it as make test/test_unix_connect.rb

So I decided to add some debug output:


  def setup
...
puts "\n\nsetup"
system 'ls -lR /tmp/kgio*'
  end

  def teardown
puts "\nteardown"
system 'ls -lR /tmp/kgio*'
...
end

  def test_unix_socket_new_invalid
puts "\ntest in"
system 'ls -lR /tmp/kgio*'
    assert_raises(ArgumentError) { Kgio::UNIXSocket.new('*' * 1024 * 1024) }
puts "\ntest out"
system 'ls -lR /tmp/kgio*'
  end

And voila:

# Running tests:
...
setup
/tmp/kgio_unix_120130906-13330-m89438:
total 0
srwxr-xr-x 1 root root 0 Sep  6 01:23 kgio_unix_120130906-13330-ch4t73

test in
/tmp/kgio_unix_120130906-13330-m89438:
total 0
srwxr-xr-x 1 root root 0 Sep  6 01:23 kgio_unix_120130906-13330-ch4t73

test out
/tmp/kgio_unix_120130906-13330-m89438:
total 0

teardown
/tmp/kgio_unix_120130906-13330-m89438:
total 0
E
...

The socket has disappeared somewhere. Interesting, isn't it? The more
interesting that kfreebsd on amd64 doesn't have this issue. Both are
up-to-date Debian Sid.

Any ideas?

> I think the problem is caused by parallel tests invocation in ruby 1.9

It seems that I was wrong: it's possible to invoke tests in parallel,
and there is an explicit option for this `--jobs'. But it doesn't
mean, that tests are invoked in parallel by default.