From: KOSAKI Motohiro <kosaki.motohiro@...>
Date: 2011-04-12T21:31:46+09:00
Subject: [ruby-core:35725] Re: [Ruby 1.9 - Bug #4558][Assigned] TestSocket#test_closed_read fails after r31230

2011/4/12  <redmine@ruby-lang.org>:
>
> Issue #4558 has been updated by Usaku NAKAMURA.
>
> Category changed from core to test
> Status changed from Closed to Assigned
> Assignee set to Motohiro KOSAKI
>
> I have no opinion about this topic, but the test code which was checked in at r31260 by kosaki-san is platform dependent.
> It blocks on Windows, and stops all tests.
>
> I request to revert it.
> Or, please explain grounds from which this test should be accepted as behavior of ruby.

I succuseeded to reporoduce this issue. On win32, IO.close() cause hang-up.
So, I think we have to discuss two thing.
 1) Why close() makes hang-up? Is it acceptable behavior?
 2) At [ruby-core:35203], We decided IO.close() raise exception to
othread threads
     and then they should wake up as ruby-1.8.
     Should we think win32 is exception for this rule?

In the other words, We have to decided rb_thread_io_blocking_region()
spec. Otherwise
we can't make any testcase. ;-)