Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

CPANtesters failing to report distribution name for Net-SSLeay

0 views
Skip to first unread message

James E Keenan

unread,
Jan 9, 2024, 4:45:05 PM1/9/24
to cpan-teste...@perl.org, h...@radiatorsoftware.com
CPANtesters appears to be failing to record the name of at least one
distribution when receiving reports. The distribution which I have
observed this in is Net-SSLeay, whose version 1.94 was uploaded on Jan
08 2024 (https://metacpan.org/dist/Net-SSLeay). If I go to the matrix
(http://matrix.cpantesters.org/?dist=Net-SSLeay or
http://fast-matrix.cpantesters.org/?dist=Net-SSLeay), however, the
*displayed* version is 1.93_05, and I also see:

#####
Other versions
NOTE: no report for latest version 1.94
#####

So the system is aware that 1.94 has been released to CPAN but believes
that no reports have yet been uploaded.

This discrepancy was called to my attention by Net-SSLeay's maintainers
in the course of resolving an issue:
https://github.com/radiator-software/p5-net-ssleay/issues/425#issuecomment-1882944459.

I then tried to install Net::SSLeay against a perl-5.39.6 where
Task::CPAN::Reporter had been installed and was known to work properly.
The report was generated as expected:

#####
[bullseye:.cpanreporter] 2011 $ ls -l reports-sent.db
-rw-r--r-- 1 jkeenan jkeenan 42304 Jan 9 20:39 reports-sent.db
[bullseye:.cpanreporter] 2012 $ tail -n 5 reports-sent.db
test PASS MooseX-Has-Sugar-1.000006 (perl-5.39.6) x86_64-linux
5.10.0-25-amd64
test PASS Email-Address-XS-1.05 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS File-MMagic-1.30 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS Modern-Perl-1.20230106 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS Net-SSLeay-1.94 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
#####

As a precaution, I then generated CPANtesters reports for one of my own
distributions on the same machine:

#####
CPAN::Reporter: Test result is 'pass', All tests successful.
CPAN::Reporter: preparing a CPAN Testers report for Pod-Multi-0.10
CPAN::Reporter: sending test report with 'pass' via Metabase
JKEENAN/Pod-Multi-0.10.tar.gz
/usr/bin/make test -- OK
#####

#####
$ tail -n 5 reports-sent.db
test PASS Net-SSLeay-1.94 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS String-Random-0.32 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS File-Save-Home-0.11 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS IO-Capture-0.05 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
test PASS Pod-Multi-0.10 (perl-5.39.6) x86_64-linux 5.10.0-25-amd64
#####

When I checked http://metabase.cpantesters.org/tail/log.txt, I found
normal entries for the last 4 distributions listed above ...

#####
[2024-01-09T21:09:46Z] [James E Keenan] [pass]
[JKEENAN/Pod-Multi-0.10.tar.gz] [x86_64-linux] [perl-v5.39.6]
[6a9cba0c-af33-11ee-b8ad-c2f41c0db3cd] [2024-01-09T21:09:46Z]
[2024-01-09T21:09:39Z] [James E Keenan] [pass]
[REYNOLDS/IO-Capture-0.05.tar.gz] [x86_64-linux] [perl-v5.39.6]
[66b57988-af33-11ee-97d0-48f41c0db3cd] [2024-01-09T21:09:39Z]
[2024-01-09T21:09:33Z] [James E Keenan] [pass]
[JKEENAN/File-Save-Home-0.11.tar.gz] [x86_64-linux] [perl-v5.39.6]
[62f380ce-af33-11ee-8d94-caf31c0db3cd] [2024-01-09T21:09:33Z]
#####

... but (what I believe to be) the entry for Net-SSLeay was defective:

#####
[2024-01-09T20:39:38Z] [James E Keenan] [pass] [] [x86_64-linux]
[perl-v5.39.6] [356cdcc6-af2f-11ee-ad9c-81f21c0db3cd] [2024-01-09T20:39:38Z]
#####

Note the '[]' after '[pass]'. I see 6 such defective entries now at
http://metabase.cpantesters.org/tail/log.txt, all from reputable
CPANtester contributors.

Can you investigate?

Thank you very much.
James E Keenan

Slaven Rezic

unread,
Jan 11, 2024, 4:30:04 PM1/11/24
to cpan-teste...@perl.org

09. 01. 2024. u 22:38, James E Keenan piše:

CPANtesters appears to be failing to record the name of at least one distribution when receiving reports.  The distribution which I have observed this in is Net-SSLeay, whose version 1.94 was uploaded on Jan 08 2024 (https://metacpan.org/dist/Net-SSLeay).  If I go to the matrix (http://matrix.cpantesters.org/?dist=Net-SSLeay or http://fast-matrix.cpantesters.org/?dist=Net-SSLeay), however, the *displayed* version is 1.93_05, and I also see:

[...]

Same problem for SignalWire-RestAPI-1.5
The matrix knows that this version exists, but there are only reports for 1.4 available.

-- Slaven

James E Keenan

unread,
Jan 11, 2024, 5:30:04 PM1/11/24
to cpan-teste...@perl.org
I don't see anything wrong with the META.json for either distro. (The
"name" element there is where I would expect CPANtesters to pick up a
distro's name.)

James E Keenan

unread,
Jan 12, 2024, 3:15:04 PM1/12/24
to cpan-teste...@perl.org
On 1/11/24 17:19, James E Keenan wrote:
> On 1/11/24 16:16, Slaven Rezic wrote:
>> 09. 01. 2024. u 22:38, James E Keenan piše:
>>
>>> CPANtesters appears to be failing to record the name of at least one
>>> distribution when receiving reports.  The distribution which I have
>>> observed this in is Net-SSLeay, whose version 1.94 was uploaded on Jan
>>> 08 2024 (https://metacpan.org/dist/Net-SSLeay). If I go to the matrix
>>> (http://matrix.cpantesters.org/?dist=Net-SSLeay or
>>> http://fast-matrix.cpantesters.org/?dist=Net-SSLeay), however, the
>>> *displayed* version is 1.93_05, and I also see:
>>
>> [...]
>>
>> Same problem for SignalWire-RestAPI-1.5
>> The matrix knows that this version exists, but there are only reports
>> for 1.4 available.
>>
>> -- Slaven
>>

Whom would we contact to investigate this further? Or, where should we
report this?

Slaven Rezic

unread,
Feb 21, 2024, 7:45:05 AM2/21/24
to cpan-teste...@perl.org

12. 01. 2024. u 21:10, James E Keenan piše:

I opened an issue after an analysis: https://github.com/cpan-testers/cpantesters-backend/issues/27

-- Slaven

James E Keenan

unread,
Feb 21, 2024, 9:00:06 AM2/21/24
to Slaven Rezic, cpan-teste...@perl.org
Thanks for your investigation. To put it mildly, our infrastructure
faces major challenges.
0 new messages