Skip Menu |

This queue is for tickets about the Net-SSLeay CPAN distribution.

Report information
The Basics
Id: 132425
Status: resolved
Priority: 0/
Queue: Net-SSLeay

People
Owner: chrisn [...] cpan.org
Requestors: PLICEASE [...] cpan.org
SREZIC [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: 1.89_03



Subject: tests fail / hang on fresh Ubuntu 20.04 install
Here is part of the test oputput: t/local/39_pkcs12.t .................... ok t/local/40_npn_support.t ............... Use of uninitialized value in string eq at t/local/40_npn_support.t line 93. Use of uninitialized value in numeric eq (==) at t/local/40_npn_support.t line 94. t/local/40_npn_support.t ............... 1/7 # Failed test 'P_next_proto_negotiated/server' # at t/local/40_npn_support.t line 59. # got: 'spdy/2' # expected: undef # Failed test 'ssl_read_all compare' # at t/local/40_npn_support.t line 62. # got: '' # expected: 'ssleay-npn-test' # Failed test 'P_next_proto_negotiated/client' # at /usr/share/perl/5.30/Test/Builder.pm line 152. # Failed test 'P_next_proto_last_status/client' # at /usr/share/perl/5.30/Test/Builder.pm line 152. # Looks like you failed 2 tests of 7. t/local/40_npn_support.t ............... Dubious, test returned 2 (wstat 512, 0x200) Failed 4/7 subtests t/local/41_alpn_support.t .............. Use of uninitialized value in string eq at t/local/41_alpn_support.t line 92. t/local/41_alpn_support.t .............. 1/6 # Failed test 'P_alpn_selected/server' # at t/local/41_alpn_support.t line 58. # got: undef # expected: 'spdy/2' # Failed test 'ssl_read_all compare' # at t/local/41_alpn_support.t line 61. # got: '' # expected: 'ssleay-alpn-test' # Failed test 'P_alpn_selected/client' # at /usr/share/perl/5.30/Test/Builder.pm line 152. # Looks like you failed 1 test of 6. t/local/41_alpn_support.t .............. Dubious, test returned 1 (wstat 256, 0x100) Failed 3/6 subtests t/local/42_info_callback.t ............. 1/2 # Failed test 'ctx: CB_HANDSHAKE_START CB_CONNECT_EXIT' # at t/local/42_info_callback.t line 100. # Failed test 'ssl: CB_HANDSHAKE_START CB_CONNECT_EXIT' # at t/local/42_info_callback.t line 100. # Looks like you failed 2 tests of 2. t/local/42_info_callback.t ............. Dubious, test returned 2 (wstat 512, 0x200) Failed 2/2 subtests t/local/43_misc_functions.t ............ ok t/local/44_sess.t ...................... Use of uninitialized value $end in string eq at t/local/44_sess.t line 219. t/local/44_sess.t ...................... Failed 58/58 subtests t/local/45_exporter.t .................. Tested both with the system perl 5.30.0 and by building 5.31.10 from source. patches from here: http://archive.ubuntu.com/ubuntu/pool/main/libn/libnet-ssleay-perl/libnet-ssleay-perl_1.88-2ubuntu1.debian.tar.xz specifically debian/patches/tlsv1.2min.patch fixes it, although is probably not (without some modifications) appropriate for other systems. Test machine happens to be a Raspberry Pi running 64bit, but I don't think that is relevant.
On Sun Apr 26 09:21:13 2020, PLICEASE wrote: Show quoted text
> Test machine happens to be a Raspberry Pi running 64bit, but I don't > think that is relevant.
Verified that I get the same fail on x86_64 ubuntu 20.04
Thanks for the report, Graham. On Sun Apr 26 14:21:13 2020, PLICEASE wrote: Show quoted text
> patches from here: > > http://archive.ubuntu.com/ubuntu/pool/main/libn/libnet-ssleay- > perl/libnet-ssleay-perl_1.88-2ubuntu1.debian.tar.xz > > specifically debian/patches/tlsv1.2min.patch fixes it, although is > probably not (without some modifications) appropriate for other > systems.
While this will probably only benefit distros that have set their default OpenSSL security level to 2 (which is most of the major ones by now), it'd be sensible to merge it anyway - there shouldn't be any downside in doing so for those still using security level 1. I'll check with Dimitri if he's happy for us to take this upstream.
On 2020-04-26 09:21:13, PLICEASE wrote: Show quoted text
> Here is part of the test oputput: > > t/local/39_pkcs12.t .................... ok > t/local/40_npn_support.t ............... Use of uninitialized value in > string eq at t/local/40_npn_support.t line 93. > Use of uninitialized value in numeric eq (==) at > t/local/40_npn_support.t line 94. > t/local/40_npn_support.t ............... 1/7 > # Failed test 'P_next_proto_negotiated/server' > # at t/local/40_npn_support.t line 59. > # got: 'spdy/2' > # expected: undef > > # Failed test 'ssl_read_all compare' > # at t/local/40_npn_support.t line 62. > # got: '' > # expected: 'ssleay-npn-test' > > # Failed test 'P_next_proto_negotiated/client' > # at /usr/share/perl/5.30/Test/Builder.pm line 152. > > # Failed test 'P_next_proto_last_status/client' > # at /usr/share/perl/5.30/Test/Builder.pm line 152. > # Looks like you failed 2 tests of 7. > t/local/40_npn_support.t ............... Dubious, test returned 2 > (wstat 512, 0x200) > Failed 4/7 subtests > t/local/41_alpn_support.t .............. Use of uninitialized value in > string eq at t/local/41_alpn_support.t line 92. > t/local/41_alpn_support.t .............. 1/6 > # Failed test 'P_alpn_selected/server' > # at t/local/41_alpn_support.t line 58. > # got: undef > # expected: 'spdy/2' > > # Failed test 'ssl_read_all compare' > # at t/local/41_alpn_support.t line 61. > # got: '' > # expected: 'ssleay-alpn-test' > > # Failed test 'P_alpn_selected/client' > # at /usr/share/perl/5.30/Test/Builder.pm line 152. > # Looks like you failed 1 test of 6. > t/local/41_alpn_support.t .............. Dubious, test returned 1 > (wstat 256, 0x100) > Failed 3/6 subtests > t/local/42_info_callback.t ............. 1/2 > # Failed test 'ctx: CB_HANDSHAKE_START CB_CONNECT_EXIT' > # at t/local/42_info_callback.t line 100. > > # Failed test 'ssl: CB_HANDSHAKE_START CB_CONNECT_EXIT' > # at t/local/42_info_callback.t line 100. > # Looks like you failed 2 tests of 2. > t/local/42_info_callback.t ............. Dubious, test returned 2 > (wstat 512, 0x200) > Failed 2/2 subtests > t/local/43_misc_functions.t ............ ok > t/local/44_sess.t ...................... Use of uninitialized value > $end in string eq at t/local/44_sess.t line 219. > t/local/44_sess.t ...................... Failed 58/58 subtests > t/local/45_exporter.t .................. > > Tested both with the system perl 5.30.0 and by building 5.31.10 from > source. > > patches from here: > > http://archive.ubuntu.com/ubuntu/pool/main/libn/libnet-ssleay- > perl/libnet-ssleay-perl_1.88-2ubuntu1.debian.tar.xz > > specifically debian/patches/tlsv1.2min.patch fixes it, although is > probably not (without some modifications) appropriate for other > systems. > > Test machine happens to be a Raspberry Pi running 64bit, but I don't > think that is relevant.
CPAN.pm users can now use this distroprefs file for automatic patching: https://github.com/eserte/srezic-cpan-distroprefs/blob/master/Net-SSLeay.yml#L90-L99
I'm running the OEM kernel with Ubuntu 20.04 which is currently Linux hostname 5.6.0-1018-oem #18-Ubuntu SMP Fri Jun 26 08:01:26 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux
Adding |5.6.0-1018-oem #18-ubuntu to myuname fixed it for me, thanks for the distroprefs file, didn't know the feature before.
On 2020-07-23 08:01:00, ABRAXXA wrote: Show quoted text
> Adding |5.6.0-1018-oem #18-ubuntu to myuname fixed it for me, thanks > for the distroprefs file, didn't know the feature before.
The test suite started to fail on debian/bullseye, so I added more myuname values.
Any plans on a clean release? This is a little headache when using docker, and cpanm to install stuff, as distroprefs are not that easy to apply. Thanks
This has been fixed in the following PRs: https://github.com/radiator-software/p5-net-ssleay/pull/212 https://github.com/radiator-software/p5-net-ssleay/pull/225 I've just released developer version 1.89_03, which includes these changes, on CPAN. If the feedback from the CPAN testers is good, our intention is to promote this to stable release 1.90 next weekend. Apologies for taking so long to fix this - it required a lot more work than we initially anticipated, because dropping to security level 1 in each of the affected tests caused failures in other environments. We've substantially restructured the test suite and generated a new PKI for it so that it runs under OpenSSL security level 2, including Ubuntu's custom security level 2 (which forbids the use of anything below TLSv1.2).