Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • Working with collection sprints
  • Document (rather than refactor) for feature specs
    • Pull them out into a separate route
    • Linked to Documentation WG
    • Created an issue in response to these needs
    • Easier to publish a page outlining the feature spec. refactoring proposal
    • Collaborated with Giarlo and Rayle

Rayle

  • Basic proposal is that the feature tests woiuld be extracted out of main test suiteOwn 
  • Own run in CI, run weekly or monthly
    • Wouldn't block builds for each Pull Request
  • The proposal details this approach

Hyrax 2.1 Release Update

Van Tuyl

  • First round of QA testing will finish this week

Diaz

  • Tests in script have been executed on at least one browser and on O/S
  • Continuing to work through this process later today
  • Aim for a beta release by next week

Van Tuyl

  • 14 blocking issues to the release exist
  • Call for collaboration on resolving these issues

Rayle

  • 8 are being actively resolved

Van Tuyl

  • If you've done work on an issue related to this project, you might be contacted
  • Any questions?

Robinson

  • Files that a Work contains when Work is public but the Files are private
    • Blocks the file name (changed to "File"), but still renders that the file is there
    • Thumbnail is also hidden, but is rendered in IIIF Viewer
    • Clicking on the link directs to an "Unauthorized" page
    • Is that the desired behavior?
    • Not a bug from the Collections Extensions Sprint

Rayle

  • Was it intentional to render those private Files to unauthenticated users?

Robinson

  • Can extend a SearchBuilder to filter by access controls
  • Should this be undertaken?

Van Tuyl

  • Interested in contacting the Repository Managers IG

Rayle

  • Should also contact the Tech. List (samvera-tech)
  • Two use cases
    • Shouldn't render the files at all
  • Perhaps permit configuration to render or hide these Files?

Hyrax 2.0 Stable Support

Johnson

  • Two weeks, the 2.0 stable build has been broken
  • 5 individual dependency shifts caused some problem
    • SimpleForm breaks, for example
    • Fixes were not backported
  • Johnson has been addressing technical debt in order to ensure that 2.0 stable continues to be stable when these updates are applied
  • How best to address this collectively?
  • https://github.com/samvera/hyrax/pull/2765
    • Was merged this morning by Colvard
    • 2.0 is now, once again, stable
  • What commitment do we have to stable releases?

Van Tuyl

  • Has the community had a policy or standard around this?
  • If not, what does everyone see as the best practices

Colvard

  • No established practices
  • Adopters are expected to undertake the backporting
  • Certain institutions might want to stay with 2.0 build given the issues involved in supporting dependency updates

Giarlo

  • Demand for maintenance on existing builds has far exceeded the resources available to undertake this work

Johnson

  • Many of the solutions required more time to sort through the commit history
  • Ideally, this time could be reduced significantly through improving reviewing of commits

Lindner

  • Code Stability WG is attempting to address these types of problems
  • Beyond simply regression tests, how can one diagnose a particular problem

Myers

  • Might this be related to Gemspecs with too loosely-versioned Gem dependencies

Johnson

  • No, for example, SimpleForm had a bugfix which broke hydra-forms functionality

Myers

  • But, when moving to a stable release, if a Gemspec versions were tighter, perhaps these problems could be mitigated
  • Not advocating that every Gem be pinned down for a patch release, but this should be addressed for a stable release

Rayle

  • Addressing changes in dependencies by pinning releases can be done
  • But more broadly, labeling which commits/pull requests resolve specific bugs might resolve this more effectively

Gondron

  • Perhaps could research which releases are affected by a given bug before one starts to resolve the bug

Rayle

  • That's ideal, but requires a system in place to support back-releases
  • Requiring one to analyze other releases before resolving a bug might further reduce the number of potential contributors

Giarlo

  • Governance WG has released recommendations around how this could be better organized
  • Resourcing
    • Finding adequate resourcing is essential
    • Forming a new WG to develop a model for better resourcing (particularly for addressing maintenance)
    • Invitation for others to become involved

Colvard

  • Feedback for the findings of the Governance WG is due today

Van Tuyl

  • Is there a more immediate need to deal with backporting issues before process is defined

Johnson

  • It could be fine right now
  • Triaging before a bug is resolved sounds fine
  • Finding which Hyrax versions are supported, and which features are being backported
  • If you are doing code review, you might be looking for this type of difference
    • Answering that for supported Hyrax releases would be an improvement

Van Tuyl

  • Added this topic to the next Hyrax Roadmap Advisory IG meeting
  • Any other issues can come to the tech. call as they arise

Moderator:

Lindner

Notetaker:

Colvard

(Hyrax Pull Request Grooming Follows)