Bug #38476 for IO-Async: New build break in IO::Async-0.16

Thu Aug 14 08:40:50 2008 david [...] davidfavor.com - Ticket created

Subject:	New build break in IO::Async-0.16
Date:	Thu, 14 Aug 2008 07:40:30 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

t/52loop-listen.................Test died early - listen actually succeeded # Looks like you planned 10 tests but only ran 9. # Looks like your test died just after 9. Dubious, test returned 255 (wstat 65280, 0xff00) Failed 1/10 subtests t/60sequencer-client............ok If you require more information email me. -- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Thu Aug 14 17:34:06 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Thu Aug 14 08:40:50 2008, david@davidfavor.com wrote: Show quoted text

> t/52loop-listen.................Test died early - listen actually

succeeded Show quoted text

> # Looks like you planned 10 tests but only ran 9. > # Looks like your test died just after 9. > Dubious, test returned 255 (wstat 65280, 0xff00) > Failed 1/10 subtests > t/60sequencer-client............ok > > If you require more information email me.

That's quite interesting. That particular test is the one at the end of the script, which tries to listen on tcp/22 or tcp/80, hoping that at least one will fail, either from being not-root or having something else listening there. For it to succeed, you'd have to be root, and not running a web or ssh server, _and_ the previous test to see if those ports are available would have to fail. Can you let me know please * What OS * If you're running the test script as root [if applicable] * If you're running a web server on port 80 or an SSH server on port 22 In the meantime, you could try just adding the port number to the message; like this: --- t/52loop-listen.t 2008-06-17 15:33:47 +0000 +++ t/52loop-listen.t 2008-08-14 21:33:28 +0000 @@ -124,7 +124,7 @@ on_resolve_error => sub { die "Test died early - resolve error $_[0]\n"; }, - on_listen => sub { die "Test died early - listen actually succeeded\n"; }, + on_listen => sub { die "Test died early - listen on port $badport actually succeeded\n"; }, on_accept => sub { "DUMMY" }, # really hope this doesn't happen ;) and run the test again, see which port it claims. -- Paul Evans

Thu Aug 14 17:34:08 2008 The RT System itself - Status changed from 'new' to 'open'

Thu Aug 14 17:55:56 2008 david [...] davidfavor.com - Correspondence added

Subject:	Re: [rt.cpan.org #38476] New build break in IO::Async-0.16
Date:	Thu, 14 Aug 2008 16:55:03 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

Paul Evans via RT wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=38476 > > > On Thu Aug 14 08:40:50 2008, david@davidfavor.com wrote:

>> t/52loop-listen.................Test died early - listen actually

> succeeded

>> # Looks like you planned 10 tests but only ran 9. >> # Looks like your test died just after 9. >> Dubious, test returned 255 (wstat 65280, 0xff00) >> Failed 1/10 subtests >> t/60sequencer-client............ok >> >> If you require more information email me.

> > That's quite interesting. That particular test is the one at the end of > the script, which tries to listen on tcp/22 or tcp/80, hoping that at > least one will fail, either from being not-root or having something else > listening there. For it to succeed, you'd have to be root, and not > running a web or ssh server, _and_ the previous test to see if those > ports are available would have to fail. > > Can you let me know please > > * What OS

net1#uname -a Linux net1.coolsurf.com 2.6.24.4-64.fc8 #1 SMP Sat Mar 29 09:54:46 EDT 2008 i686 i686 i386 GNU/Linux Show quoted text

> * If you're running the test script as root [if applicable]

Yes. Show quoted text

> * If you're running a web server on port 80 or an SSH server on port 22

I have a Web server running on port 80. I just moved my sshd to an odd port to reduce hack attempts. If I move sshd back to port 22 the test works. Move it to another port and the test fails. Show quoted text

> In the meantime, you could try just adding the port number to the > message; like this: > > --- t/52loop-listen.t 2008-06-17 15:33:47 +0000 > +++ t/52loop-listen.t 2008-08-14 21:33:28 +0000 > @@ -124,7 +124,7 @@ > > on_resolve_error => sub { die "Test died early - resolve error > $_[0]\n"; }, > > - on_listen => sub { die "Test died early - listen actually > succeeded\n"; }, > + on_listen => sub { die "Test died early - listen on port $badport > actually succeeded\n"; }, > > on_accept => sub { "DUMMY" }, # really hope this doesn't happen ;) > > and run the test again, see which port it claims. >

-- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Thu Aug 14 19:57:01 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Thu Aug 14 17:55:56 2008, david@davidfavor.com wrote: Show quoted text

> > * If you're running the test script as root [if applicable]

> > Yes. >

> > * If you're running a web server on port 80 or an SSH server on

> port 22 > > I have a Web server running on port 80. > > I just moved my sshd to an odd port to reduce hack attempts. > > If I move sshd back to port 22 the test works. Move it to another > port and the test fails.

Ah yes. I can confirm if I run it as root, with ssh stopped, this indeed happens. I'll have a fiddle with it and make it work. -- Paul Evans

Thu Aug 14 20:00:23 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

[the attached] patch seems to fix it for me. -- Paul Evans

=== modified file 't/52loop-listen.t' --- t/52loop-listen.t 2008-08-14 21:42:57 +0000 +++ t/52loop-listen.t 2008-08-14 23:58:44 +0000 @@ -108,7 +108,12 @@ # it's likely we'll fail to bind TCP port 22 or 80. my $badport; -IO::Socket::INET->new( Type => SOCK_STREAM, LocalPort => $_, Listen => 1 ) or $badport = $_, last for 22, 80; +foreach my $port ( 22, 80 ) { + IO::Socket::INET->new( Type => SOCK_STREAM, LocalPort => $port, Listen => 1, ReuseAddr => 1 ) and next; + + $badport = $port; + last; +} SKIP: { skip "No bind()-failing ports found", 1 unless defined $badport;

Thu Aug 14 21:57:12 2008 david [...] davidfavor.com - Correspondence added

Subject:	Re: [rt.cpan.org #38476] New build break in IO::Async-0.16
Date:	Thu, 14 Aug 2008 20:56:55 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

Paul Evans via RT wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=38476 > > > On Thu Aug 14 17:55:56 2008, david@davidfavor.com wrote:

>>> * If you're running the test script as root [if applicable]

>> Yes. >>

>>> * If you're running a web server on port 80 or an SSH server on

>> port 22 >> >> I have a Web server running on port 80. >> >> I just moved my sshd to an odd port to reduce hack attempts. >> >> If I move sshd back to port 22 the test works. Move it to another >> port and the test fails.

> > Ah yes. I can confirm if I run it as root, with ssh stopped, this indeed > happens. I'll have a fiddle with it and make it work. >

Great. I'll test the new version as soon as it shows up on CPAN. -- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Thu Aug 14 21:57:57 2008 david [...] davidfavor.com - Correspondence added

Subject:	Re: [rt.cpan.org #38476] New build break in IO::Async-0.16
Date:	Thu, 14 Aug 2008 20:57:46 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

Paul Evans via RT wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=38476 > > > [the attached] patch seems to fix it for me. > >

I'll just wait for the CPAN version as it pulls to my local mirror and rebuilds each morning. -- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Mon Aug 18 19:54:10 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Thu Aug 14 21:57:57 2008, david@davidfavor.com wrote: Show quoted text

> I'll just wait for the CPAN version as it pulls to my local > mirror and rebuilds each morning.

OK. Let me know how you get on with: http://search.cpan.org/~pevans/IO-Async-0.16.001/ and if it's good then I'll call it 0.17 -- Paul Evans

Tue Aug 19 09:14:14 2008 david [...] davidfavor.com - Correspondence added

Subject:	Re: [rt.cpan.org #38476] New build break in IO::Async-0.16
Date:	Tue, 19 Aug 2008 08:13:55 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

Paul Evans via RT wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=38476 > > > On Thu Aug 14 21:57:57 2008, david@davidfavor.com wrote:

>> I'll just wait for the CPAN version as it pulls to my local >> mirror and rebuilds each morning.

> > OK. Let me know how you get on with: > > http://search.cpan.org/~pevans/IO-Async-0.16.001/ > > and if it's good then I'll call it 0.17 >

Perfect! All tests succeed. Thanks for fixing this. -- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Wed Sep 10 03:04:34 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Tue Aug 19 09:14:14 2008, david@davidfavor.com wrote: Show quoted text

> Perfect! All tests succeed. > > Thanks for fixing this.

OK. Closing bug. -- Paul Evans

Wed Sep 10 03:04:37 2008 leonerd-cpan [...] leonerd.org.uk - Status changed from 'open' to 'resolved'

Wed Sep 10 03:05:04 2008 leonerd-cpan [...] leonerd.org.uk - Fixed in 0.17 added

Wed Sep 10 12:10:19 2008 david [...] davidfavor.com - Correspondence added

Subject:	Re: [rt.cpan.org #38476] New build break in IO::Async-0.16
Date:	Wed, 10 Sep 2008 11:10:03 -0500
To:	bug-IO-Async [...] rt.cpan.org
From:	David Favor <david [...] davidfavor.com>

Paul Evans via RT wrote: Show quoted text

> <URL: http://rt.cpan.org/Ticket/Display.html?id=38476 > > > On Tue Aug 19 09:14:14 2008, david@davidfavor.com wrote:

>> Perfect! All tests succeed. >> >> Thanks for fixing this.

> > OK. Closing bug. > >

Thanks. -- Love feeling your best ever, all day, every day? Click http://RadicalHealth.com/join for the easy way.

Wed Sep 10 12:10:22 2008 The RT System itself - Status changed from 'resolved' to 'open'

Sun Nov 23 08:15:16 2008 leonerd-cpan [...] leonerd.org.uk - Correspondence added

Closing. Again -- Paul Evans

Sun Nov 23 08:15:46 2008 leonerd-cpan [...] leonerd.org.uk - Status changed from 'open' to 'resolved'