Bug #100578 for IO-Socket-IP: make test fails in t/16v6only.t on FreeBSD 10.1 stable (r274698)

Thu Nov 27 06:31:08 2014 frank [...] undermydesk.org - Ticket created

Subject:	make test fails in t/16v6only.t on FreeBSD 10.1 stable (r274698)
Date:	Thu, 27 Nov 2014 12:30:47 +0100
To:	bug-IO-Socket-IP [...] rt.cpan.org
From:	Frank Reppin <frank [...] undermydesk.org>

affected IO-Socket-IP version: ================================== IO-Socket-IP-0.33.tar.gz affected perl version: ================================== ... see attachment 01 affected OS: ================================== FreeBSD amazone.undermydesk.org 10.1-STABLE FreeBSD 10.1-STABLE #0 r274698: Wed Nov 19 13:00:46 CET 2014 frank@undermydesk.org:/usr/obj/usr/src/sys/AMAZONE amd64 failed message is: ================================== # Failed test 'Socket creation fails with connection refused' # at t/16v6only.t line 43. # 'Operation timed out' # doesn't match '(?^:Connection\ refused)' # Looks like you failed 1 test of 6. t/16v6only.t ........................ Dubious, test returned 1 (wstat 256, 0x100) Failed 1/6 subtests full build log (attachment 02) is attached. additional info: ================================== Had some minor troubles brewing 5.20.1 as well - they all seem to be somehow IPv6 related (IPv6 is enabled here but not in use): Test Summary Report ------------------- ../cpan/IO-Socket-IP/t/16v6only.t (Wstat: 256 Tests: 6 Failed: 1) Failed test: 3 Non-zero exit status: 1 ../dist/Net-Ping/t/450_service.t (Wstat: 512 Tests: 26 Failed: 2) Failed tests: 9, 18 Non-zero exit status: 2 ../dist/Net-Ping/t/510_ping_udp.t (Wstat: 256 Tests: 2 Failed: 1) Failed test: 2 Non-zero exit status: 1 Files=2411, Tests=710379, 2642 wallclock secs (299.29 usr 20.80 sys + 2032.48 cusr 152.92 csys = 2505.48 CPU) Result: FAIL *** Error code 4 Stop. make: stopped in /opt/perl5/build/perl-5.20.1 Please let me know if I you need anything else to get this sorted. TIA, frank\ -- 43rd Law of Computing: Anything that can go wr fortune: Segmentation violation -- Core dumped

Message body is not shown because sender requested not to inline it.

Fri Nov 28 14:15:56 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-11-27 06:31:08, frank@undermydesk.org wrote: Show quoted text

> affected IO-Socket-IP version: > ================================== > > IO-Socket-IP-0.33.tar.gz > > affected perl version: > ================================== > > ... see attachment 01 > > affected OS: > ================================== > > FreeBSD amazone.undermydesk.org 10.1-STABLE FreeBSD 10.1-STABLE #0 > r274698: Wed Nov 19 13:00:46 CET 2014 > frank@undermydesk.org:/usr/obj/usr/src/sys/AMAZONE amd64 > > > failed message is: > ================================== > > # Failed test 'Socket creation fails with connection refused' > # at t/16v6only.t line 43. > # 'Operation timed out' > # doesn't match '(?^:Connection\ refused)' > # Looks like you failed 1 test of 6. > t/16v6only.t ........................ > Dubious, test returned 1 (wstat 256, 0x100) > Failed 1/6 subtests > > > full build log (attachment 02) is attached. > > additional info: > ================================== > > Had some minor troubles brewing 5.20.1 as well - they > all seem to be somehow IPv6 related (IPv6 is enabled here > but not in use): > > Test Summary Report > ------------------- > ../cpan/IO-Socket-IP/t/16v6only.t (Wstat: > 256 Tests: 6 Failed: 1) > Failed test: 3 > Non-zero exit status: 1 > ../dist/Net-Ping/t/450_service.t (Wstat: > 512 Tests: 26 Failed: 2) > Failed tests: 9, 18 > Non-zero exit status: 2 > ../dist/Net-Ping/t/510_ping_udp.t (Wstat: > 256 Tests: 2 Failed: 1) > Failed test: 2 > Non-zero exit status: 1 > Files=2411, Tests=710379, 2642 wallclock secs (299.29 usr 20.80 sys + > 2032.48 cusr 152.92 csys = 2505.48 CPU) > Result: FAIL > *** Error code 4 > > Stop. > make: stopped in /opt/perl5/build/perl-5.20.1 > > > > Please let me know if I you need anything else to > get this sorted. > > TIA, > frank\ >

It seems that there are FreeBSD 10 systems where the test suite passes. See http://matrix.cpantesters.org/?dist=IO-Socket-IP%200.33;os=freebsd;reports=1#sl=2,1 Regards, Slaven

Fri Nov 28 14:15:56 2014 The RT System itself - Status changed from 'new' to 'open'

Tue Dec 02 18:56:22 2014 frank [...] undermydesk.org - Correspondence added

Subject:	[rt.cpan.org #100578] make test fails in t/16v6only.t on FreeBSD 10.1 stable (r274698)
Date:	Wed, 03 Dec 2014 00:56:00 +0100
To:	bug-IO-Socket-IP [...] rt.cpan.org
From:	Frank Reppin <frank [...] undermydesk.org>

Hellos, yeah - false positive... my apologies! Please close this issue. The reason why it failed for me was to have net.inet.tcp.blackhole set to non zero (default value is 0) as I later noticed: [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: sysctl net.inet.tcp.blackhole=0 net.inet.tcp.blackhole: 1 -> 0 [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: perl t/16v6only.t ok 1 - IPV6_V6ONLY is 1 on $listensock ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true ok 3 - Socket creation fails with connection refused ok 4 - IPV6_V6ONLY is 0 on $listensock ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false ok 6 - Test socket connected to correct peer port 1..6 [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: sysctl net.inet.tcp.blackhole=1 net.inet.tcp.blackhole: 0 -> 1 [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: perl t/16v6only.t ok 1 - IPV6_V6ONLY is 1 on $listensock ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true not ok 3 - Socket creation fails with connection refused # Failed test 'Socket creation fails with connection refused' # at t/16v6only.t line 43. # 'Operation timed out' # doesn't match '(?^:Connection\ refused)' ok 4 - IPV6_V6ONLY is 0 on $listensock ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false ok 6 - Test socket connected to correct peer port 1..6 # Looks like you failed 1 test of 6. [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: sysctl net.inet.tcp.blackhole=0 net.inet.tcp.blackhole: 1 -> 0 [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: perl t/16v6only.t ok 1 - IPV6_V6ONLY is 1 on $listensock ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true ok 3 - Socket creation fails with connection refused ok 4 - IPV6_V6ONLY is 0 on $listensock ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false ok 6 - Test socket connected to correct peer port 1..6 cheers, Frank Reppin -- 43rd Law of Computing: Anything that can go wr fortune: Segmentation violation -- Core dumped

Wed Dec 03 01:28:35 2014 SREZIC [...] cpan.org - Correspondence added

On 2014-12-02 18:56:22, frank@undermydesk.org wrote: Show quoted text

> Hellos, > > yeah - false positive... my apologies! > Please close this issue. > > The reason why it failed for me was to have > > net.inet.tcp.blackhole > > set to non zero (default value is 0) as I later noticed: > > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > sysctl net.inet.tcp.blackhole=0 > net.inet.tcp.blackhole: 1 -> 0 > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > perl t/16v6only.t > ok 1 - IPV6_V6ONLY is 1 on $listensock > ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true > ok 3 - Socket creation fails with connection refused > ok 4 - IPV6_V6ONLY is 0 on $listensock > ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false > ok 6 - Test socket connected to correct peer port > 1..6 > > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > sysctl net.inet.tcp.blackhole=1 > net.inet.tcp.blackhole: 0 -> 1 > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > perl t/16v6only.t > ok 1 - IPV6_V6ONLY is 1 on $listensock > ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true > not ok 3 - Socket creation fails with connection refused > # Failed test 'Socket creation fails with connection refused' > # at t/16v6only.t line 43. > # 'Operation timed out' > # doesn't match '(?^:Connection\ refused)' > ok 4 - IPV6_V6ONLY is 0 on $listensock > ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false > ok 6 - Test socket connected to correct peer port > 1..6 > # Looks like you failed 1 test of 6. > > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > sysctl net.inet.tcp.blackhole=0 > net.inet.tcp.blackhole: 1 -> 0 > [frank@amazone]/usr/ports/net/p5-IO-Socket-IP/work/IO-Socket-IP-0.33: > perl t/16v6only.t > ok 1 - IPV6_V6ONLY is 1 on $listensock > ok 2 - Unable to connect PF_INET socket to PF_INET6 socket with V6Only true > ok 3 - Socket creation fails with connection refused > ok 4 - IPV6_V6ONLY is 0 on $listensock > ok 5 - Connected PF_INET socket to PF_INET6 socket with V6Only false > ok 6 - Test socket connected to correct peer port > 1..6 >

Maybe the test can be adapted to check for the net.inet.tcp.blackhole value, and skip the test or allow for a timeout error in this case?

Sun Jan 11 12:00:43 2015 leonerd-cpan [...] leonerd.org.uk - Correspondence added

On Wed Dec 03 01:28:35 2014, SREZIC wrote: Show quoted text

> Maybe the test can be adapted to check for the net.inet.tcp.blackhole > value, and skip the test or allow for a timeout error in this case?

Not easily. That value is specific to OpenBSD, so the test would have to work around similar trickery on other OSes. It wouldn't be very portable. This sort of thing is always difficult to write into unit tests... given a small low-level wrapper around OS functionality, like IO::Socket::IP is, it's hard to write a test just of that wrapper without effectively also testing the underlying OS it sits on. This test fails because the OS doesn't allow such packets to pass, which is entirely the OS and not the wrapper code's fault. Short of just deleting the test entirely, I can't see what else can be done. -- Paul Evans

Tue Aug 02 13:13:58 2016 leonerd-cpan [...] leonerd.org.uk - Correspondence added

Closing as requested -- Paul Evans

Tue Aug 02 13:13:58 2016 leonerd-cpan [...] leonerd.org.uk - Status changed from 'open' to 'resolved'