[Pulp-dev] RPM plugin meeting notes

[Tatiana Tereshchenko]

Pulp 2:

   -

   Multi-repo copy update (dalley)
   -

      PRs ready for review
      -

      https://github.com/pulp/pulp_rpm/pull/1420/
      -

      https://github.com/pulp/pulp/pull/3947
      -

      Docs changes remain
      -

      Still no smash tests, will work with QE on this, they won’t have
      capacity until upcoming Wednesday
      -

      Need to clear simple copy changes with all stakeholders?  Katello is
      already +1


Pulp 3:

   -

   Kickstart update (david/fabricio)
   -

      Splitted in 3 PRs:
      -

         Models: https://github.com/pulp/pulp_rpm/pull/1418
         -

         Sync: https://github.com/pulp/pulp_rpm/pull/1427/
         <https://github.com/pulp/pulp_rpm/pull/1427/files>


   -

         Publish: https://github.com/pulp/pulp_rpm/pull/1429
         -

      Tested sync on RHEL7/RHEL8 repos
      -

      Sync seems to be working, needing some feedback to improve it, and to
      understand publishing
      -

         Suggestions: add test of Centos7/8 repos, manual test of
         multi-variant KS from RHEL8 ContentISO, add testing of on-demand



   -

   Modularity update (ppicka)
   -

      Sync in progress
      -

      Changes to pipeline to handle artifacts without remote artifacts
      - check the pulp-dev thread and Matthias' approach
         -

         Discussion: modulemd always exists, because metadata is always
         downloaded even in the deferred-download scenario



   -

   Comps.xml update (dawalker/ttereshc)
   -

      Name collision - rpm_group, had to rename Group to PackageGroup, any
      other suggestions?
      -

         Rename other entities (eg PackageEnvironment) to be consistent?
         - or rename rpm_* fields in Package model?
         - AI: think
         -

      Do we want to have any relations like group->packages, group->
      categories, etc?
      -

         It seems like in pulp 2 we have a recursive copy for categories?
         https://github.com/pulp/pulp_rpm/blob/2-master/plugins/pulp_rpm/plugins/importers/yum/associate.py#L112
         -

         Katello doesn’t need recursive copy for any comps.xml types
         -

         Discussion: implementing it is complicated, but not doing it could
         cause us way more grief later - consensus seems to be “get the model
         correct from the start”
         -

   copy API
   -

      suggestion: new endpoint that does recursive-copy and mark
      simple-copy as “you can, but probably shouldn’t”
      -

      take discussion to list


Open PRs:

   -

   https://github.com/pulp/pulp_rpm/pull
   <https://github.com/pulp/pulp_rpm/pulls>
   - ppicka to rebase module upload PR


Triage:

   -

   Un-triaged bugs https://pulp.plan.io/projects/pulp_rpm/issues?query_id=30
   -

   Triage etherpad https://etherpad.net/p/rpm_triage_grooming
   -

      https://pulp.plan.io/issues/5367
      - new item requested by Katello (it's a pulp 2 feature but it was not
         requested before)
         -

         bmbouter to add details and convert to a story
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-dev/attachments/20190829/f0319288/attachment.htm>