#opnfv-meeting: dovetail weekly meeting

Meeting started by hongbo at 14:00:24 UTC (full logs).

Meeting summary

    1. roll name (hongbo, 14:01:26)
    2. Hongbo Tian (hongbo, 14:01:36)
    3. leo (leo_wang, 14:02:22)
    4. https://wiki.opnfv.org/display/dovetail/Candidate+Dovetail+test+use+cases (hongbo, 14:04:23)

  1. Requirement that all patches be submitted upstream (hongbo, 14:05:39)
    1. Chris said we raise that to the c&C meeting (hongbo, 14:08:44)
    2. dave we start continue to the conversation of this, and to the C&C (hongbo, 14:12:53)
    3. Dave proposes that we agree to recommend that the OPNFV reference platform should not require patches which have not been accepted upstream to be able to pass the Dovetail test suite (dneary, 14:22:08)
    4. Dave gives two examples: SFC requires an NSH patch to OVS which is not upstream, and the RT_PREEMPT patches which are not completely integrated into the kernel. So real-time kernel or NSH enabled OVS should not be required to pass Dovetail tests. (dneary, 14:24:01)
    5. Chris says that "upstream" needs clearer definition, since many new projects can be considered "upstreams" (dneary, 14:24:28)
    6. Dave asks whether we need to maintain a list of upstream projects and their reference tree (dneary, 14:25:15)
    7. Dave proposes "Dovetail tests may not rely on patches which have not been accepted into the main source tree of an open source project required as a dependency of an OPNFV scenario" (dneary, 14:28:29)
    8. We have moved on to debating scenarios, and whether it is appropriate to require that multiple stacks/scenarios must support a feature before we add it to Dovetail (dneary, 14:42:22)
    9. Chris says, if one project supports upstream a feature which is important to us, and a competing community is not interested, or does not accept a patch enabling it (dneary, 14:44:13)
    10. Dave argues that it would be seen as a political statement in favour of one project over another. Chris argues that refusing to add a feature will be seen as a political statement too (dneary, 14:47:08)
    11. Chris has issues with limiting feature verification to features present in multiple stacks. Dave responds that exercising features is not the role of Dovetail. (dneary, 14:48:41)
    12. Wenjing says there is general consensus on requiring that patches be accepted upstream, and asks if we need to add a loophole to allow exceptions (dneary, 14:50:10)
    13. Dave says that we should not add an exception rule, but if in the future the Dovetail committers agree to accept an exception, we will have the authority to do that. (dneary, 14:51:34)
    14. Zenghui agrees that patches should be upstream, and Dovetail can decide which scenarios to target for Dovetail test case validation (dneary, 14:53:28)
    15. We agreed that the "Out of scope" section can be removed from the requirements document (dneary, 14:57:13)
    16. We agreed to defer discussion of multiple scenario question due to lack of consensus (dneary, 14:57:38)
    17. We agreed to include an upstream requirement in the test case requirements (dneary, 14:57:57)
    18. On test case documentation, Chris proposed to review the documentation requirements to make it clearer what his expectations for test cases are (dneary, 14:58:33)
    19. Wenjing brought up the need to be more rigorous about recording meeting agendas and minutes, and proposed that we need to re-organize the top-level of the wiki (dneary, 14:59:22)
    20. Dave agrees to help with the wiki re-organization (dneary, 15:00:35)

Meeting ended at 15:00:42 UTC (full logs).

Action items

  1. (none)

People present (lines said)

  1. dneary (22)
  2. hongbo (9)
  3. collabot (3)
  4. leo_wang (2)
  5. Wenjing (2)

Generated by MeetBot 0.1.4.