FHIR Chat · Newbie Issues · IG creation

Stream: IG creation

Topic: Newbie Issues

Richard Kavanagh (Jun 24 2019 at 09:44):

Hi - attempting to run the publisher for the first time on windows.

Jekyll is instaled.
The publisher was downloaded from https://fhir.github.io/latest-ig-publisher/org.hl7.fhir.publisher.jar

Just to test things out I have cloned the master branch of IPS from https://github.com/HL7/fhir-ips

I run the publisher in GUI mode, selected the ig.json file and clicked execute.

The tool is busy for a little while but when it completes it has not created any output (though it does create the output folder).
I can see it has downloaded the .fhir\packages\hl7.fhir.core#4.0.0 package

What am I doing wrong?

Grahame Grieve (Jun 24 2019 at 10:38):

what's the output?

Grahame Grieve (Jun 24 2019 at 10:38):

console output

Lynn Laakso (Jun 24 2019 at 12:44):

I have similar problems with it in GUI mode. have reverted to command line mode

Richard Kavanagh (Jun 24 2019 at 12:53):

The UI output is pasted image

Lynn Laakso (Jun 24 2019 at 12:54):

my UI output runs off the window and I never see it complete

Lynn Laakso (Jun 24 2019 at 12:54):

don't have a solution just using command line

Grahame Grieve (Jun 24 2019 at 13:02):

I don't know why it would get stuck there

Richard Kavanagh (Jun 24 2019 at 13:55):

Looks like the issue is a "Heap" issue - I get much further with the command line. Looks like I am going to need more memory in my machine though ...

Richard Kavanagh (Jun 24 2019 at 13:55):

As an aside in IPS there is an issue as er pasted image

Richard Kavanagh (Jun 24 2019 at 13:56):

presumably caused by pasted image

Grahame Grieve (Jun 24 2019 at 20:44):

@James Agnew this is an issue in the hapi types. The relevant code in question is

      if (e.hasPrimitiveValue())
        x.addText(((DateTimeType) e).toHumanDisplay());

Grahame Grieve (Jun 24 2019 at 20:45):

in this case, there is an extension and no value, so hasPrimitiveValue() should return false... but the code says it will return true because it's a primitive type. That feels like a bug to me, but it also feels ..difficult... to fix?

James Agnew (Jun 25 2019 at 21:03):

@Grahame Grieve This seems straightforward enough to me... All primitives extend PrimitiveType, so we'd probably want something like the following in PrimitiveType:

  @Override
  public boolean hasPrimitiveValue() {
    return StringUtils.isNotBlank(getValueAsString());
  }

Grahame Grieve (Jun 25 2019 at 21:09):

thats what I think - is it a breaking change?

Grahame Grieve (Jun 25 2019 at 21:10):

what we don't have is anywhere to make release notes - where should they go?

James Agnew (Jun 25 2019 at 21:48):

I can document that in HAPI FHIR's changelog ( http://hapifhir.io/changes-report.html )- I don't know that I'd consider this a breaking change so much as a bug fix, since this is clearly correcting the behaviour to match the name of the method.

And either way, august release will be 4.0.0 with a bunch of other minor breaking changes in other parts of the API (server mostly) so this is probably the time for it.

Grahame Grieve (Jun 25 2019 at 21:49):

ok. well, I'll commit the change to R3-R5 and 2016May sometime today. please add a note to the change log

James Agnew (Jun 25 2019 at 21:54):

Done.

Last updated: Apr 12 2022 at 19:14 UTC