- ensure that when validating codes against value sets, the valid coding is actually returned
- strip out the new extensions we add for FMM
- improve the rendering of Quantity and Range. (The former was horrid and the latter wasn't as nice as it could be.)
## Validator Changes:
* ignore canonical resources found in core examples packages
* get .ofType() working in discriminators (round #1!)
* fix bug checking enableWhen - ignoring items in answers
* Improved Error messages validating bundle entries
## Other Code Changes
* Improve rendering of uris that point to known resources
* Fix wrong reference rendering questionnaire
* Fix rendering of QuestionnaireResponses - render items in answers properly
* Improve Error message when snapshot can't be generated
* fix up support for R5
***NO_CI***
## Validator
* Allow both -output and -output-style parameters (output-style applies to output now) + add csv output style
* Implement the ```-level``` parameter
* Add support for R4B validation
* fix bug in deep profiles (profiles that don't start at the root)
## Other code changes
* Regenerate R5 code for 5.0.0-snapshot1 & many consequential changes
* Add version support for R4B and R5 (R4B code generation still to be done)
***NO_CI***
* Allow both -output and -output-style parameters (output-style applies to output now) + add csv output style
* Implement the ```-level``` parameter
* Add support for R4B validation
* fix bug in deep profiles (profiles that don't start at the root)
## Other code changes
* Regenerate R5 code for 5.0.0-snapshot1 & many consequential changes
* Add version support for R4B and R5 (R4B code generation still to be done)
Validator:
* check StructureDefinition derivation consistency
* fix bug looking for contained resources inside bundles
* fix up determination criteria for IRIs on identifiers in urn:ietf:rfc:3986
Other code changes:
* Add support for new FHIR releases
* Fix escaping bug in graphQL encoder
***NO_CI***
1 out of 500 runs would produce a 0, meaning that _JUST_UNDER or _JUST_OVER values would be actually be _EXACT, which would fail by definition.
This adds a minimum value of 1 to nextInt()
Lloyd McKenzie
Grahame Grieve
markiantorno
jamesagnew
dotasek
dotasek