Stream: implementers
Topic: Bundle Invariants
Brian Postlethwaite (Dec 15 2016 at 01:10):
Just wondering if a bundle being returned from a resource's history should also satisfy all the invariants, as I don't think that invariant bdl-7 can be satisfied if there are deletes in the results (as they don't have a resource, and therefore no meta.versionId).
Brian Postlethwaite (Dec 15 2016 at 01:10):
This is causing some of the connectathon tests in Touchstone to fail. (on my server anyway)
Grahame Grieve (Dec 15 2016 at 01:12):
multiple deletes on the same resource?
Brian Postlethwaite (Dec 15 2016 at 01:27):
can do.
Brian Postlethwaite (Dec 15 2016 at 01:27):
a create, delete, update (which re-creates) then delete would do it
Richard Ettema (Dec 15 2016 at 05:58):
What if bld-7 and bld-8 only applied when (type = 'history).not()?
Then add a new invariant for when (type = 'history) where fullUrl is version specific.
Brian Postlethwaite (Dec 15 2016 at 06:01):
bdl-8 I'm ok with, bdl-7 I think needs your addition in there.
Richard Ettema (Dec 15 2016 at 06:07):
So bld-7 would look like this:
(type = 'history').not() and entry.where(fullUrl).select(fullUrl&resource.meta.versionId).isDistinct()
Richard Ettema (Dec 15 2016 at 06:12):
If that looks right I can update Touchstone's fhir validator for 1.8.0. (Tomorrow - heading off to bed now)
Brian Postlethwaite (Dec 15 2016 at 06:14):
Cool, now that I've updated my FhirPathTester to support both DSTU2 and STU3, should be easy to check.
Richard Ettema (Dec 15 2016 at 06:16):
Nice. I'll let you know when I have that done.
(Signing off now...)
Richard Ettema (Dec 16 2016 at 01:17):
@Brian Postlethwaite I've updated the bdl-7 invariant in Touchstone and ran one of the TestScripts that had the failure and it now passes.
Richard Ettema (Dec 16 2016 at 01:19):
The actual change to bdl-7 was slightly different. It's now defined as:
(type = 'history') or entry.where(fullUrl).select(fullUrl&resource.meta.versionId).isDistinct()
Last updated: Apr 12 2022 at 19:14 UTC
