Skip Menu |

This queue is for tickets about the LWP-Protocol-https CPAN distribution.

Report information
The Basics
Id: 67001
Status: resolved
Priority: 0/
Queue: LWP-Protocol-https
People
Owner: Nobody in particular
Requestors: blue [...] thisisnotmyrealemail.com
Cc:
AdminCc:
Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)

History Show all quoted text
  Sun Mar 27 12:24:59 2011 blue [...] thisisnotmyrealemail.com - Ticket created
Subject: t/apache.t fails: connection refused
apache.org apparently is refusing connection based on useragent, so the test fails. i swapped apache.org with bankofamerica.com and it worked. it might be a good idea to cycle through a randomized list of sites, and pass the test on the first successful site. also, maybe this should be added as the first test: ok($ua->is_protocol_supported('https'));
  Sat Feb 18 17:53:20 2012 GAAS [...] cpan.org - Correspondence added
Seems unlikely that apache.org is denying access based on that. Must be some other reason.
  Sat Feb 18 17:53:20 2012 The RT System itself - Status changed from 'new' to 'open'
  Wed Dec 26 16:16:38 2012 cpan [...] barely3am.com - Correspondence added
i think this is an ipv6 problem related to: http://stackoverflow.com/questions/11463748/perl-iosocketssl-connect-network-is- unreachable i was [finally] able to re-produce in CentOS6 x64, even with (i think) ipv6 turned off in the kernel, the app is trying to leverage the ipv6 nameserver response (for some reason it seems like RHEL wants to load up the ipv6 mod, but leave it 'disabled', which might be screwing up other code when they get ipv6 records back from NS and then repsond back with a network failure) i tested against apache and google.com, same thing, added the 'inet4' to a fork of t/apache.t and it seems to work again. this doesn't seem to happen on debian or ubuntu.. (fwiw), so i'm guessing it's a RHEL derivative thing. i dunno that it's "safe" to have these sorts of network tests w/o explicit v4 or v6 assignments... even though it's correct in saying something's wrong, wget and other apps seem to work OK, so maybe this is something that needs to be solved further down in the stack...(?) probably a RHEL / IO::Socket::SSL bug (or similar), but for now, would it be safe to implicitly put inet4 in the test script? something like: use IO::Socket::SSL 'inet4';
  Thu Dec 27 06:50:08 2012 cpan [...] barely3am.com - Correspondence added
more data: after drilling down a bit more, i think it's a system call issue with centos and io-socket-inet6.. still something to be aware of with the tests (imo).
  Thu Dec 27 07:08:40 2012 cpan [...] barely3am.com - Correspondence added
confirmed, bug is in io-socket-inet6 < 2.56 or so. seems to be fixed by 2.69, problem is the author of inet6 has some test bugs that need resolving. a straight install works and appears to solve the problem. i checked on ubuntu and that version of INET6.pm was > than 2.55, but that machine also has full ipv6 routing too... again, something to be aware of when writing network tests...
  Sun Mar 02 14:10:26 2014 dsteinbrunner [...] pobox.com - Correspondence added
From: dsteinbrunner [...] pobox.com
On Thu Dec 27 07:08:40 2012, SAXJAZMAN wrote: Show quoted text
> confirmed, bug is in io-socket-inet6 < 2.56 or so. seems to be fixed > by 2.69, problem is the > author of inet6 has some test bugs that need resolving. a straight > install works and appears to > solve the problem. i checked on ubuntu and that version of INET6.pm > was > than 2.55, but that > machine also has full ipv6 routing too... > > again, something to be aware of when writing network tests...
I ran into this issue trying to install Crypt::SSLeay on Mac OS X 10.6 with perl 5.10. Crypt::SSLeay was suggesting that libssl-dev was missing but after some digging and finding this I updated IO::Socket::INET6 and Crypt::SSLeay along with LWP::Protocol::https installed properly.
  Sun Mar 02 14:25:31 2014 dsteinbrunner [...] pobox.com - Correspondence added
From: dsteinbrunner [...] pobox.com
On Sun Mar 02 14:10:26 2014, dsteinbrunner@pobox.com wrote: Show quoted text
> I updated > IO::Socket::INET6 and Crypt::SSLeay along with LWP::Protocol::https > installed properly.
note that it was updated from version 2.56 to 2.72
  Mon Mar 03 23:19:03 2014 dsteinbrunner [...] pobox.com - Correspondence added
From: dsteinbrunner [...] pobox.com
On Sun Mar 02 14:25:31 2014, dsteinbrunner@pobox.com wrote: Show quoted text
> On Sun Mar 02 14:10:26 2014, dsteinbrunner@pobox.com wrote: >
> > I updated > > IO::Socket::INET6 and Crypt::SSLeay along with LWP::Protocol::https > > installed properly.
> > note that it was updated from version 2.56 to 2.72
I just tried using 2.57 and it did not allow LWP::Protocol::https to install but 2.58 did.
  Fri Mar 07 10:15:24 2014 dsteinbrunner [...] pobox.com - Correspondence added
From: dsteinbrunner [...] pobox.com
On Mon Mar 03 23:19:03 2014, dsteinbrunner@pobox.com wrote: Show quoted text
> I just tried using 2.57 and it did not allow LWP::Protocol::https to > install but 2.58 did.
Reported this as an issue with IO::Socket::INET6 here: https://rt.cpan.org/Public/Bug/Display.html?id=93503 The next release should make things happy
  Fri Mar 31 15:32:36 2017 olaf [...] wundersolutions.com - Correspondence added
Ticket migrated to github as https://github.com/libwww-perl/LWP-Protocol-https/issues/44
  Fri Mar 31 15:32:42 2017 olaf [...] wundersolutions.com - Status changed from 'open' to 'resolved'