3-1-stable: AR XML serialization tests broken for anyone else?

47 views
Skip to first unread message

Steve Jorgensen

unread,
Sep 14, 2012, 12:39:08 AM9/14/12
to rubyonra...@googlegroups.com
Hi all, I'm preparing a pull request for 3-1-stable, but I'm getting test failures in the XML serialization tests (running just AR tests for postgresql). I looked on Travis, but I can't see that we have builds for release branches. If they are there, please advise how to find those.

I haven't dug into the cause yet to see if it's complicated or obvious, but here are the 6 failures I'm seeing...

  1) Failure:
test_should_serialize_binary(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:124]:
Failed assertion, no message given.

  2) Failure:
test_should_serialize_boolean(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:139]:
Failed assertion, no message given.

  3) Failure:
test_should_serialize_datetime(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:132]:
Failed assertion, no message given.

  4) Failure:
test_should_serialize_integer(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:117]:
Failed assertion, no message given.

  5) Failure:
test_should_serialize_string(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:113]:
Expected /<name nil="true"><\/name>/ to match "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<xml-contact>\n  <name nil=\"true\"/>\n  <age type=\"integer\" nil=\"true\"/>\n  <avatar type=\"binary\" nil=\"true\" encoding=\"base64\"/>\n  <created-at type=\"datetime\" nil=\"true\"/>\n  <awesome type=\"boolean\" nil=\"true\"/>\n  <preferences nil=\"true\"/>\n</xml-contact>\n".

  6) Failure:
test_should_serialize_yaml(NilXmlSerializationTest) [/Users/FileVault/stevej/Projects/rails/activerecord/test/cases/xml_serialization_test.rb:146]:
Expected /<preferences nil=\"true\"><\/preferences>/ to match "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n<xml-contact>\n  <name nil=\"true\"/>\n  <age type=\"integer\" nil=\"true\"/>\n  <avatar type=\"binary\" nil=\"true\" encoding=\"base64\"/>\n  <created-at type=\"datetime\" nil=\"true\"/>\n  <awesome type=\"boolean\" nil=\"true\"/>\n  <preferences nil=\"true\"/>\n</xml-contact>\n".

Steve Klabnik

unread,
Sep 14, 2012, 1:02:30 AM9/14/12
to rubyonra...@googlegroups.com
3-1 isn't getting any non-security related bug fixes these days.

Steve Jorgensen

unread,
Sep 14, 2012, 1:07:27 AM9/14/12
to rubyonra...@googlegroups.com
Oh -- good to know. If I submit a trivial non-security pull request, is it likely to be accepted?

Jon Leighton

unread,
Sep 14, 2012, 5:11:01 AM9/14/12
to rubyonra...@googlegroups.com
On 14/09/12 06:07, Steve Jorgensen wrote:
> Oh -- good to know. If I submit a trivial non-security pull request, is
> it likely to be accepted?

If you submit a PR against 3.1 with the sole purpose of fixing broken
tests, I'll happily merge. We'll need those tests to work if/when there
are future security patches to apply anyway.

--
http://jonathanleighton.com/

Rafael Mendonça França

unread,
Sep 14, 2012, 9:58:01 AM9/14/12
to rubyonra...@googlegroups.com
Security patch should not be sent in the issues tracker. Read this page if you think your bug is a security one http://rubyonrails.org/security

--
You received this message because you are subscribed to the Google Groups "Ruby on Rails: Core" group.
To post to this group, send email to rubyonra...@googlegroups.com.
To unsubscribe from this group, send email to rubyonrails-co...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/rubyonrails-core?hl=en.


Steve Jorgensen

unread,
Sep 20, 2012, 3:51:08 PM9/20/12
to rubyonra...@googlegroups.com
I had been wanting to back-port a PostgreSQL auto-reconnect fix. Since that's not a security issue though, I guess it's not appropriate for Rails 3.1. I have written a trivial initializer for my employer that patches the problem in 3.1, so I'll just ask permission to publish that as a Gist for anyone who needs it.

I'll probably still go ahead and fix the build if no one beats (or has beaten) me to it.
Reply all
Reply to author
Forward
0 new messages