Stream: IG creation
Topic: issue on canonical urls in 4.6.0
François Macary (Jul 04 2021 at 10:30):
I'm migrating the order-catalog IG from R4 to R5 (in a new branch in github) and have hundrededs of errors regarding canonical urls mentioned in profiles of the IG. For instance:
image.png
with a strange repetition of the version part of the canonical url, which also appears in the profile itself:
image.png
I'm wondering if this might come from the template for 4.6.0.
Thanks in advance for any advices.
Lloyd McKenzie (Jul 04 2021 at 15:25):
There aren't version-specific templates. Is 4.6.0 appearing in your source anywhere other than the ImplementationGuide.fhirVersion?
François Macary (Jul 04 2021 at 17:26):
Only there and in the package-list.jason file
Lloyd McKenzie (Jul 04 2021 at 18:27):
Is there a branch committed that Mark might test with?
François Macary (Jul 05 2021 at 07:48):
Yes, the branch name is catalogR5: https://github.com/HL7/fhir-order-catalog/tree/catalogR5
Grahame Grieve (Jul 05 2021 at 11:34):
I thought I fixed this. It's an error in the 4.6.0 package itself
François Macary (Aug 13 2021 at 11:56):
I built the IG with the latest publisher today and still have the problem.
So, what would be the fix? Is there something I must do on my side?
In summary, I have some 500 errors of the kind above. And I have also other hundreds of broken links saying:
StructureDefinition-InputObservationDefinition.profile.ttl.html#/html/body/div/div/div/p/a at Line 3454, column 240 error The link 'http://build.fhir.org/' for "FHIR 4.6.0" cannot be resolved, as in the snapshot below:
image.png
Other examples that make me think of a problem of initialization of base url:
image.png
image.png
Lloyd McKenzie (Aug 13 2021 at 14:10):
There does seem to be an issue with adjusting URLs of inherited text to refer to the base of the inherited spec rather than the current spec. There's already a Git issue about this
François Macary (Sep 10 2021 at 14:53):
@Lloyd McKenzie do you have an idea of when an action will be taken to fix this issue?
Lloyd McKenzie (Sep 10 2021 at 18:17):
@Grahame Grieve
Grahame Grieve (Sep 14 2021 at 20:37):
4.6.0 is http://hl7.org/fhir/2021May/, so it's an error refer to build.fhir.org
Grahame Grieve (Sep 14 2021 at 20:37):
@François Macary
François Macary (Sep 15 2021 at 07:22):
Thanks. I'll try that. I guess it means that the profiles of the IG will apply to the base resources from the current build.
Grahame Grieve (Sep 15 2021 at 10:32):
from the last publication
François Macary (Sep 17 2021 at 08:28):
I'm sorry, I'm not sure I captured what I need to do. What should I put into :
- package-list.json : element fhirversion (where version = "current") ?
- the IG.fhirversion ?
Should I replace "4.6.0" by "current" in both elements?
Rob Hausam (Sep 17 2021 at 12:02):
@Grahame Grieve How should this be specified? At least previously, 'current' didn't work?
John Moehrke (Sep 17 2021 at 12:40):
would be nice if this value could be just "4", or "4.6", or "4.0"...
John Moehrke (Sep 17 2021 at 12:42):
@François Macary I think you are wanting what is in the CapabilityStatement.rest.resource.profile
François Macary (Sep 17 2021 at 13:12):
Could you point me to it? The order-catalog IG does not have such a profile until now. I see the element in the base standard of course. But what I'm trying to do is to solve hundreds of resolution errors in my FHIR 4.6.0 IG.
François Macary (Sep 17 2021 at 13:18):
And I confirm that fhirversion = "current" does not work. It's probably not what @Grahame Grieve meant.
Rob Hausam (Sep 17 2021 at 13:19):
Yes. It's just not clear yet what the correct way to resolve this is.
Grahame Grieve (Sep 17 2021 at 20:26):
perhaps I wasn't clear enough: #current is not supported
Rob Hausam (Sep 18 2021 at 01:57):
The real question is, how should @François Macary deal with the issue(s) with the IG that he is asking about?
@Grahame Grieve
Grahame Grieve (Sep 18 2021 at 03:55):
I do not know. I really don't. There's no road map for this at the moment
Last updated: Apr 12 2022 at 19:14 UTC
