Stream: fmg
Topic: Vote required - temporarily waive warnings - Approved
Lloyd McKenzie (Mar 19 2017 at 15:41):
Somehow, at the last minute, we're seeing warnings pop up on a few infrastructure resources all dealing with search criteria that aren't covered by examples. This is due to either late-breaking changes to examples that removed elements from examples or possibly changes to the code that detects what examples are covered. I'm not sure why and don't have the time to investigate. You can see the issues in http://build.fhir.org/qa.html (look for numbers in the red column). My recommendation is to add these warnings to the "temporarily waived" set of warnings so they do not impede the ability to publish these resources with their correct FMM level. We will address them for the next release.
Lloyd McKenzie (Mar 19 2017 at 15:41):
+1
Brian Pech (Mar 19 2017 at 15:44):
Agree +1
Brian Postlethwaite (Mar 19 2017 at 15:52):
+1 (if its needed)
John Moehrke (Mar 19 2017 at 15:56):
+1 (Im willing, but worried about the 'somehow' clause)
Paul Knapp (Mar 19 2017 at 16:14):
The totals by work group, if that is what they are, are not correct.
Paul Knapp (Mar 19 2017 at 16:17):
Several of these are are FMM 5 resources. I think it would be better to fix them pronto - don't we have another rule in place which is that we won't make changes to FMM 5 resources pre normative ballot?
Lloyd McKenzie (Mar 19 2017 at 16:21):
@Paul Knapp The rule is that we'll consult the implementer community before making substantive changes to FMM4+ resources. Adding examples would not be substantive, so there's no issue there. The only impact of this not being fixed is that it's slightly harder for those building the reference implementations to test all of their search criteria.
Lloyd McKenzie (Mar 19 2017 at 16:22):
@John Moehrke If you've got time to dig through the logs or roll-back until you see when the problem was introduced, please do. Agree it would be good to know.
David Hay (Mar 19 2017 at 16:41):
+1
Lloyd McKenzie (Mar 19 2017 at 16:46):
Approved. I will hold off making the change in case Grahame wants to investigate the underlying cause. Thanks for your quick turn-around everyone. (And if you can check your email frequently today on the off chance we run into other issues requiring decision, that'd be appreciated.)
John Moehrke (Mar 19 2017 at 18:10):
the searchParameter warning is legitimate. there is a search parameter for which it doesn't have an example with that element filled. This seems to have happened as far back as December. I do not get the warnings on StructureDefinition, so...StructureDefinition is not anything I understand, so not much help from me.
Hans Buitendijk (Mar 19 2017 at 18:12):
+1
Grahame Grieve (Mar 19 2017 at 19:19):
I sure haven
Grahame Grieve (Mar 19 2017 at 19:20):
I sure haven't made changes in this space. what's an example?
Lloyd McKenzie (Mar 19 2017 at 19:38):
SearchParameter.component
Lloyd McKenzie (Mar 19 2017 at 19:38):
(Which we certainly *ought* to have . . .)
Grahame Grieve (Mar 19 2017 at 19:39):
so the warnings are all valid, so far as I can see. I cannot think why they're only suddenly appeared
Grahame Grieve (Mar 19 2017 at 19:40):
with regard to the conformance resources, the generated conformance resources are not tested in this regard, so the warnings on CodeSystem, OperationDefinition, SearchParameter, and StructureDefinition aren't actually true. But the others appear to be
Grahame Grieve (Mar 19 2017 at 19:41):
so +1
Lloyd McKenzie (Mar 19 2017 at 19:58):
I'm pretty sure these particular issues weren't being reported a week ago or I'd have flagged them as something we needed to fix. (I did flag SearchParameter.component, but I thought you'd addressed.) I'll mark them to ignore for now.
Last updated: Apr 12 2022 at 19:14 UTC
