Allow 6.x versions of FHIR #1517
Conversation
FHIR version 6 is currently in balloting stages. It does not appear that any code changes are needed yet for allowing an R6 IG. The usual warning about prerelease versions will be shown when one of the ballot versions is used.
There was a problem hiding this comment.
Well, that was easier than expected! I tested it out by creating an R6 IG and profiling an element that doesn't exist in R5. It worked! I was even able to run the IG Publisher after.
One of the reasons this is so simple now is that past us thought ahead to this moment and added this:
sushi/src/utils/FHIRVersionUtils.ts
Lines 61 to 75 in 04b9d7b
Anyway, nice job!
test/utils/Processing.test.ts
Outdated
| it('should allow FHIR R6', () => { | ||
| const input = path.join(__dirname, 'fixtures', 'fhir-r6'); | ||
| const config = readConfig(input); | ||
| expect(config.fhirVersion).toEqual(['6.0.0-ballot2']); | ||
| }); |
There was a problem hiding this comment.
I don't know what the difference is, but I was also able to use just 6.0.0 as a fhirVersion already. Is it worth adding a test for that? SUSHI recognizes it as a non-preRelease version, thanks to the snippet of code Chris linked.
There was a problem hiding this comment.
Getting a package downloaded for 6.0.0 right now is not so good, but either way, I can add a test for when that version is specified.
Description:
FHIR version 6 is currently in balloting stages. It does not appear that any code changes are needed yet for allowing an R6 IG. The usual warning about prerelease versions will be shown when one of the ballot versions is used.
Testing Instructions:
The main question here is whether supporting R6 is actually as simple as I think it is. Having looked over information about the R6 ballot changes, I don't think there's anything that will need special handling.
The new datatype RelativeTime is complex enough that it doesn't make sense to try to support assigning it in a single rule. But, it may be worth considering if we want to support checking its invariants. rlt-1 is a very simple check, and rlt-2 isn't too much more complicated. We do something like this for Extensions for their ext-1 invariant.
Related Issue:
#1485