Bug #6036: Test failures in Fedora Rawhide/17

• Status: Closed
• Assignee: vo.x (Vit Ondruch)
• Target version: -
• ruby -v: ruby 1.9.3p125 (2012-02-16 revision 34643) [x86_64-linux]

Hi, I'm getting these test failures [1] on both Fedora 17 and Rawhide. The tests unfortunately don't run all, because there is probably an endless loop in one of them, which I can't reproduce when running tests separately (is there a way to output each test file in 'make test', so that I can find the file with the endless loop?).
BTW I'm running tests with GDBM 1.10 and some tests are skipped claiming I have GDBM 1.8 - which is probably an issue with the regexps matching GDBM version.

Regards,
Bohuslav "Slavek" Kabrda

[1] https://gist.github.com/1845285

[akr (Akira Tanaka)]

• Status changed from Open to Closed
• % Done changed from 0 to 100

This issue was solved with changeset r34646.
Bohuslav, thank you for reporting this issue.
Your contribution to Ruby is greatly appreciated.
May Ruby be with you.

---

• test/dbm/test_dbm.rb: fix skip condition for libgdbm 1.8.0 or prior.
  reported by Bohuslav Kabrda.
  [ruby-core:42685] [Bug #6036]

Well, thanks for fixing the gdbm support, but the other 16 failures and 20 errors are the main reason why I reported the bug... Could you have a look at that?
Thanks.

[naruse (Yui NARUSE)]

• Status changed from Closed to Assigned
• Assignee set to vo.x (Vit Ondruch)

Bohuslav Kabrda wrote:

Well, thanks for fixing the gdbm support, but the other 16 failures and 20 errors are the main reason why I reported the bug... Could you have a look at that?
Thanks.

Vit, could you inspect this?

[vo.x (Vit Ondruch)]

Tests passes with following command: "make check TESTS="-v -x test_pathname.rb -x test_drbssl.rb -x test_parse.rb -x test_x509cert.rb"

The test_drbssl.rb and test_x509cert.rb might be related to the OpenSSL changes?

[vo.x (Vit Ondruch)]

The test_pathname appears to hang somewhere. The TestPathname#test_pathsubext_510 is the last output which appears in my log, however there might be some entries lost due to not flushed log.

[akr (Akira Tanaka)]

2012/2/20 Vit Ondruch [e-mail]:

The test_pathname appears to hang somewhere. The TestPathname#test_pathsubext_510 is the last output which appears in my log, however there might be some entries lost due to not flushed log.

I installed Fedora but I couldn't reproduce the problem.

% ./ruby test/pathname/test_pathname.rb
Run options:

Running tests:

...(snip)...

Finished tests in 0.520247s, 505.5290 tests/s, 1039.8904 assertions/s.

263 tests, 541 assertions, 0 failures, 0 errors, 0 skips
% ./ruby -v
ruby 1.9.3p134 (2012-02-19 revision 34690) [i686-linux]
% lsb_release -a
LSB Version: :core-4.0-ia32:core-4.0-noarch
Distributor ID: Fedora
Description: Fedora release 18 (Rawhide)
Release: 18
Codename: Rawhide

Tanaka Akira

[vo.x (Vit Ondruch)]

Akira Tanaka wrote:

2012/2/20 Vit Ondruch [e-mail]:

The test_pathname appears to hang somewhere. The TestPathname#test_pathsubext_510 is the last output which appears in my log, however there might be some entries lost due to not flushed log.

I installed Fedora but I couldn't reproduce the problem.

% ./ruby test/pathname/test_pathname.rb
Run options:

Running tests:

...(snip)...

Finished tests in 0.520247s, 505.5290 tests/s, 1039.8904 assertions/s.

263 tests, 541 assertions, 0 failures, 0 errors, 0 skips
% ./ruby -v
ruby 1.9.3p134 (2012-02-19 revision 34690) [i686-linux]
% lsb_release -a
LSB Version: :core-4.0-ia32:core-4.0-noarch
Distributor ID: Fedora
Description: Fedora release 18 (Rawhide)
Release: 18
Codename: Rawhide

Tanaka Akira

I am afraid that the test results differs if you run the whole test suite in once or if you run just part of the test suite, as you did.

[akr (Akira Tanaka)]

2012/2/22 Vit Ondruch [e-mail]:

I am afraid that the test results differs if you run the whole test suite in once or if you run just part of the test suite, as you did.

I see.

The hang is reproduced on my environment.

However the hang is occur at TestProcess#test_abort, not
TestPathname#test_pathsubext_510.

|TestProc#test_to_proc = 0.00 s = .
|TestProc#test_to_s = 0.00 s = .
|TestProcess#test_abort =
^C/home/akr/ruby/193/ruby/test/drb/ut_array_drbssl.rb:22:in join': Interrupt | from /home/akr/ruby/193/ruby/test/drb/ut_array_drbssl.rb:22:in '
|/home/akr/ruby/193/ruby/test/drb/ut_drb_drbssl.rb:23:in join': Interrupt | from /home/akr/ruby/193/ruby/test/drb/ut_drb_drbssl.rb:23:in '
|
|
|Finished tests in 3031.769106s, 2.1611 tests/s, 593.4136 assertions/s.
|
| 1) Failure:
|test_dsig_algorithm_mismatch(OpenSSL::TestX509Certificate)
[/home/akr/ruby/193/ruby/test/openssl/test_x509cert.rb:175]:
|OpenSSL::X509::CertificateError expected but nothing was raised.
|...

After ^C, bactrace shows drb.

I think the problem is caused by drb tests.

Tanaka Akira

[vo.x (Vit Ondruch)]

Akira Tanaka wrote:

2012/2/22 Vit Ondruch [e-mail]:

I am afraid that the test results differs if you run the whole test suite in once or if you run just part of the test suite, as you did.

I see.

The hang is reproduced on my environment.

However the hang is occur at TestProcess#test_abort, not
TestPathname#test_pathsubext_510.

Yes, that might be. I just noted that the "TestPathname#test_pathsubext_510" was last entry in my log, so there might be other tests passed, but not flushed yet.

|TestProc#test_to_proc = 0.00 s = .
|TestProc#test_to_s = 0.00 s = .
|TestProcess#test_abort =
^C/home/akr/ruby/193/ruby/test/drb/ut_array_drbssl.rb:22:in join': Interrupt | from /home/akr/ruby/193/ruby/test/drb/ut_array_drbssl.rb:22:in '
|/home/akr/ruby/193/ruby/test/drb/ut_drb_drbssl.rb:23:in join': Interrupt | from /home/akr/ruby/193/ruby/test/drb/ut_drb_drbssl.rb:23:in '
|
|
|Finished tests in 3031.769106s, 2.1611 tests/s, 593.4136 assertions/s.
|
| 1) Failure:
|test_dsig_algorithm_mismatch(OpenSSL::TestX509Certificate)
[/home/akr/ruby/193/ruby/test/openssl/test_x509cert.rb:175]:
|OpenSSL::X509::CertificateError expected but nothing was raised.
|...

After ^C, bactrace shows drb.

I think the problem is caused by drb tests.

Tanaka Akira

That makes sense. I now testing with drb tests disabled, and it does not hang.

[vo.x (Vit Ondruch)]

It seems that the errors are due to updated openssl library: openssl-1.0.1-0.1.beta2.fc17 ... there is a lot of errors around, Python, MySql ...

[vo.x (Vit Ondruch)]

It seems that the patch [1] changes the behavior of openssl and makes the test_x509cert.rb fail:

1. Failure:
   test_dsig_algorithm_mismatch(OpenSSL::TestX509Certificate) [test/openssl/test_x509cert.rb:175]:
   OpenSSL::X509::CertificateError expected but nothing was raised.

I also notified Fedora's openssl maintainer about these issues [2].

[1] http://cvs.openssl.org/filediff?f=openssl/crypto/asn1/a_sign.c&v1=1.21.4.1&v2=1.21.4.2
[2] https://bugzilla.redhat.com/show_bug.cgi?id=797217

[naruse (Yui NARUSE)]

Could you make another ticket ans assign it to Martin Boßlet about openssl?

[vo.x (Vit Ondruch)]

• Status changed from Assigned to Closed

Since the remaining issues were reported in separate issues #6089, #6221, I'm closing this ticket