Bug 1932203

Summary: perl-Crypt-GPG-1.64-20.fc35 FTBFS: Use of uninitialized value in string eq at t/02-import.t line 40.
Product: [Fedora] Fedora Reporter: Petr Pisar <ppisar>
Component: perl-Crypt-GPGAssignee: Robert Scheck <redhat-bugzilla>
Status: CLOSED RAWHIDE QA Contact: Fedora Extras Quality Assurance <extras-qa>
Severity: unspecified Docs Contact:
Priority: unspecified    
Version: rawhideCC: perl-devel, redhat-bugzilla
Target Milestone: ---   
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
URL: https://koschei.fedoraproject.org/package/perl-Crypt-GPG
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2021-05-25 22:30:18 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1927309    

Description Petr Pisar 2021-02-24 08:48:38 UTC
perl-Crypt-GPG-1.64-20.fc35 fails to build in mock/Koji environment because tests fail there:

+ make test
PERL_DL_NONLAZY=1 "/usr/bin/perl" "-MExtUtils::Command::MM" "-MTest::Harness" "-e" "undef *Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
NOTE: If the tests are skipped, you may need to install gpg,
      and/or set the environment variable GPGBIN to the 
      location of the gpg binary.
      If the keygen test takes a long time you may need to 
      generate more randomness on your computer (by running a 
      recursive directory listing in the background, for 
      example).
process ended prematurely at /builddir/build/BUILD/Crypt-GPG-1.64/blib/lib/Crypt/GPG.pm line 473.
process ended prematurely at /builddir/build/BUILD/Crypt-GPG-1.64/blib/lib/Crypt/GPG.pm line 473.
t/01-keygen.t .. ok
Use of uninitialized value in string eq at t/02-import.t line 40.
# Failed test 1 in t/02-import.t at line 44
t/02-import.t .. 
Failed 1/1 subtests 
[...]
Test Summary Report
-------------------
t/02-import.t (Wstat: 0 Tests: 1 Failed: 1)
  Failed test:  1
t/03-export.t (Wstat: 0 Tests: 10 Failed: 4)
  Failed tests:  1, 4, 6, 9
t/04-encdec.t (Wstat: 0 Tests: 10 Failed: 10)
  Failed tests:  1-10
t/05-sigver.t (Wstat: 2560 Tests: 1 Failed: 1)
  Failed test:  1
  Non-zero exit status: 10
  Parse errors: Bad plan.  You planned 32 tests but ran 1.
t/06-keyops.t (Wstat: 2560 Tests: 0 Failed: 0)
  Non-zero exit status: 10
  Parse errors: Bad plan.  You planned 30 tests but ran 0.

A difference between passing and failing build root is at <https://koji.fedoraproject.org/koji/taskinfo?taskID=61274500>. Because I cannot reproduces in my virtual machine, I believe this could be triggered by a new namespace-isolating mock which is also used in Koji.

Comment 1 Robert Scheck 2021-05-22 23:53:05 UTC
- Build on buildvm-x86-06.iad2.fedoraproject.org fails
- Build on buildvm-s390x-19.s390.fedoraproject.org succeeds
- Build on buildvm-s390x-20.s390.fedoraproject.org succeeds

Comment 2 Petr Pisar 2021-06-15 06:15:02 UTC
This issues still exists <https://koschei.fedoraproject.org/build/10395650>. I guess it's a non-determinist issues which clutters CI.