yahoo-eng-team team mailing list archive

[OpenStack Infra <1648054@xxxxxxxxxxxxxxxxxx>]

Reviewed:  https://review.openstack.org/408069
Committed: https://git.openstack.org/cgit/openstack/nova/commit/?id=b8c88c69f516622d92a31279161a5f228c263c39
Submitter: Jenkins
Branch:    master

commit b8c88c69f516622d92a31279161a5f228c263c39
Author: Sergey Nikitin <snikitin@xxxxxxxxxxxx>
Date:   Wed Dec 7 16:44:43 2016 +0400

    Revert "reduce pep8 requirements to just hacking"
    
    This reverts commit 1e287ae14bac519310221d187e7d1563c2310823
    
    Constraints mechanism doesn't work for 'hacking' lib:
    I11eb0d762869ad8920795fb710f1b2eeb9354f12
    
    We have to install all of test-requirements to get
    'hacking' at the right version.
    
    Closes-Bug: #1648054
    
    Change-Id: I512192e8a9ab3a53a52e6cc4d4b8a222951af7fb


** Changed in: nova
       Status: In Progress => Fix Released

-- 
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1648054

Title:
  Upper constraints for 'hacking' lib doesn't work for pep8

Status in OpenStack Compute (nova):
  Fix Released

Bug description:
  Upper constraints for 'hacking' lib doesn't work for pep8

  Couple days ago this patch was merged into Nova:
  https://review.openstack.org/#/c/334048/11/nova/tests/unit/virt/libvirt/test_vif.py

  It contains these lines:

     self.assertEqual(None, conf.vhost_queues)
     self.assertEqual(None, conf.driver_name)

  From pep8 point of view these lines are incorrect. Here 'assertIsNone'
  should be used (rule N318)

  But during merge pep8 job was 'green'. Unfortunately some number of contributors faced with fails of 'tox -e pep8' command in local repos because of these lines. It has given rise to this patch https://review.openstack.org/#/c/407870/
  This patch fixes current problem but it doesn't fix such problems in future.

  The reason of green pep8 job was in 'hacking' lib. It was released 6
  day ago and new release 0.13.0 has a bug, wich causes our problem.

  Commit with bug was already reverted from 'hacking' lib but new
  version lib without bug wasn't released yet.

  So the conclusion is: new release of 'hacking' lib breaks nova pep8 job. It becomes false positive.
  The question is "Why new release of 'hacking' was installed?"

  We have upper constraints for hacking lib in test-requirements.txt:

  hacking<0.11,>=0.10.0

  But we didn't use these constraints for pep8. For pep8 we use only
  'hacking' dependency in tox.ini without any constraints.

  https://github.com/openstack/nova/blob/master/tox.ini#L44

  Because of it we install the latest version of 'hacking' each time.

  To fix it we need to revert this patch
  https://review.openstack.org/#/c/267097/6

To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1648054/+subscriptions