2017-05-17 Design Forum

How to Join

Agenda

• Attendees
  • Wyclif
  • Darius
  • Daniel
  • Burke
  Topics
  • Creating/defining a backlog of "needs design" issues
  • https://issues.openmrs.org/browse/TRUNK-2284
  Notes
  Creating/defining a backlog of "needs design" issues
  • We have a JIRA filter, but aren't using it
  • Would like to be considering most relevant issues (not tickets that are 4 years old)
  TRUNK-2284 - Common tagging mechanism for openmrs objects (GSoC 2017)
  • Project page: https://wiki.openmrs.org/x/RoChBg
  • Proposal: tags shared across all domain objects
  • tag:
  • id
  • name
  • object_tag_map:
  • object_class
  • object_uuid
  • tag_id
  • (or, skip the mapping table, and just have object_tag have object_class, object_uuid, and tag_name)
  • Darius: my initial reaction is to skip the tag table and just have a map table.
  Burke:
      when you request a REST object (from any resource) it has "tags" on it (a list of strings) as a top-level property
      consumer API is just "add tag 'foobar' to object xyz" (you shouldn't need to create a tag first)
      would like to be able to do getHavingTag(Class<? extends OpenmrsObject>, String tag)
      
      REST API?
      GET .../tag?q=need => returns [{tag:"needs-review",count:25}]
      (a) GET .../patient?tag=needs-review
  •     => returns a list of patients
      (b) GET .../taggedobject/needs-review?t=patient
  •     => returns a bunch of links to object
  TRUNK-3158 - Explicit Control over Flush Mode
  https://github.com/openmrs/openmrs-core/blob/master/api/src/main/java/org/openmrs/api/db/hibernate/HibernateConceptDAO.java#L1840
  https://github.com/openmrs/openmrs-core/search?utf8=%E2%9C%93&q=setFlushMode&type=
  This method seems like a bug (changes to manual but never resets it): https://github.com/openmrs/openmrs-core/blob/571f3080b3d870dd7c1d3a00a01aab1fd3fe2295/api/src/main/java/org/openmrs/api/db/hibernate/HibernateOrderDAO.java#L203
  Darius: I don't like get/setFlushMode in the service. That doesn't improve things. Having a @ManualFlush method-level annotation would allow us to clean up the 7 instances of setFlushMode(MANUAL); try { ... } finally { reset } that are in openmrs-core (which is ugly code).
  
  

Notes

You can also access the notes below here.

Transcripts

• Audio recording of the call: Listen online or download (available after the meeting)