The Linux Foundation

 
Minutes Feb 2 2011

From The Linux Foundation

Attendees:

Jeff Licquia (LF), Mats Wichmann (Intel), Alan Clark (Novell), Robert Schweikert (Novell)

Agenda:

  • LSB 4.1 RC1 status. Do we need a RC2?
  • Schedule for Collab Summit.
  • If we have time, a discussion about "upstream relations".

Minutes:

Jeff: RC1 - been out for a week or so now. Status on the missing archs (s390*) - have managed to complete s390x. issues with the machine - build host loses connection between buildbot master and slave. For s390, it slow on hercules, and working out details of building 31bit on 64bit system.
Jeff: Don't need a full s390 build, we're only missing a few packages.
Jeff: IBM folks would like us to reduce the usage on the system we lost access to. We've gone back and forth on this a bit. We *may* get the system back. Unknown whether the old provisioning is still in place.
Jeff: I'm not seeing any reason why we wouldn't be ready to release all 7 archs at the end of this week or early next week.
Jeff: rundown of the bzr changes since the RC tag - apache init script, sdk gets issue, ghostscript fvt, libbat/desktop build issues wrg gets, libbat lsb version issues

Decision is to rebuild the sdk, see if we can snapshot build on the usable arches, and rebuild desktop test.

Jeff: Any other issues with regard to 4.1 RC?
Robert: Do we ant to do some cleanup of the tests, in particular the python issue with audioop?
Jeff: I think we have a handle on most of the test issues. The python issue has a spec impact. This involves works on Mats part, right?
Mats: Yes, a db change, need errata. If someone can do that, then I can do the database part
Robert: I think we need to have clean tests, where it's clear that a failure is a distribution issue and not a test issue.
Jeff: I don't we can do this for 4.1, but it's a goal to work towards
Jeff: Some of our tests are pretty clean, perhaps you can start with those
Robert: If possible yes, with the current state of the tests, I can't push the issue with our QA

Some discussion about test issues and quality of tests.

Jeff: Whether the tests are run manually or in dist-checker, a subset can be run and managed with more tests added as needed

Jeff: So for the python audioop drop, Mats will do the database work, Stew will do the errata, python-test, appchk-python bits.

Jeff: Collaboration Summit. Last week we talked about doing a 1 day early thing on Tues. I have confirmation that we'll have a room on Tues and a public session on Friday, if people want to register and make reservations.

Jeff: We've got a few minutes. I've got something else that was brought up: Upstream relations. Ideally the distributions would run the tests and treat us as upstream to file test bugs, and fix the other bugs at their end or upstream. For now, with us running the tests, how do we handle filing bugs upstream. Perhaps too big of a topic for these last few minutes of the call, but something to think about.

Jeff: I'll open the floor. Anything else?

End of Call


[Article] [Discussion] [View source] [History]