Stream: IG creation
Topic: Redirections
Grahame Grieve (Apr 10 2019 at 00:08):
I've just created all the missing canonical redirections for the IGs published on hl7.org/fhir at /us and /uv. I couldn't automatically create all of them - the record keeping I needed for them was incomplete and I'm only going to go back and fix them by request.
btw, redirection = if you access http://hl7.org/fhir/us/core/StructureDefinition/us-core-careplan (the canonical URL) you will get a 404. I'm putting the redirects for the canonical URLs in place
Grahame Grieve (Apr 10 2019 at 00:11):
please test the IGs you care about for this
Lloyd McKenzie (Apr 10 2019 at 03:04):
Should the IGPublisher not just create these automatically?
Grahame Grieve (Apr 10 2019 at 03:25):
it's not that easy; the redirection technique depends on the server software where it's going to be published. I have a special section of the IG publisher than maintains a publication set now, and it's generating the redirects. (you can tell it what the server is)
Grahame Grieve (Apr 10 2019 at 03:29):
also, I don't think people want 100s of extra files that only make sense at the canonical location
Lloyd McKenzie (Apr 10 2019 at 03:38):
Ok, then I'm not totally clear on why manual effort is required. You know the canonical URL and you know the filename so why can't the publisher just automatically generate all of them when you set the server flag to tell it what they need to look like?
Grahame Grieve (Apr 10 2019 at 03:38):
because it turns out that I don't always know those things from the published IG for the older IGs
Grahame Grieve (Apr 10 2019 at 03:39):
the publisher knew them while publishing, but the published IG isn't always clear.
Grahame Grieve (Apr 10 2019 at 03:39):
there's quite some mess in the past. Anything published since I started working on packages is fine
Grahame Grieve (Apr 10 2019 at 03:40):
and even, for some of the IGs, the files themselves are not published
Lloyd McKenzie (Apr 10 2019 at 03:40):
Ah, so the issue is the IGs that were published in the past before the IGPublisher had the code to do this. So those had to be done manually, but for everything done from this point forward, it should be automatic?
Grahame Grieve (Apr 10 2019 at 03:41):
yes from now on, it will be fine (I just overrode the fact that the author can decide whether to publish json or xml instances for HL7 and FHIRF IGs)
Lloyd McKenzie (Apr 10 2019 at 15:02):
Why?
Lloyd McKenzie (Apr 10 2019 at 15:02):
In CRD - you cannot implement XML
Lloyd McKenzie (Apr 10 2019 at 15:03):
Or is it that you need the instances, even if they're not actually visible in the IG?
Grahame Grieve (Apr 10 2019 at 20:24):
they don't have to be visible in the IG, but if you ask for [canonica] with xml as the mime type, you'll get the definition in xml
Last updated: Apr 12 2022 at 19:14 UTC
