Stream: hapi
Topic: Java Core Generator
Grahame Grieve (Dec 25 2019 at 19:04):
@James Agnew I am updating the generator for R5, and running into some challenges with some of the embedded metadata. Is it documented anywhere?
Grahame Grieve (Dec 28 2019 at 22:48):
ok, I've committed a new code generator for R5. This code generator is different in that you run it as a stand alone executable rather than as part of the main build. It just uses the published packages.
there's some documentation here: https://github.com/hapifhir/org.hl7.fhir.core/tree/master/org.hl7.fhir.core.generator
@James Agnew There's some methodology changes in R5 that impact on the generated code (the metadata parts). Dealing with them is what prompted me to finally move the code out of the build. The R5 code will need a good amount of testing; people will be looking for an R5 release prior to Sydney (ASAP, basically) so we should get it out soonish but with clear warnings: your milage may vary....
Grahame Grieve (Dec 28 2019 at 22:48):
I'll commit the new generated code for R5 shortly. It's tricky since I need to get a new build out to make the code correct, but I can't get the build updated until I have the new code released
Grahame Grieve (Dec 29 2019 at 11:38):
hmm. what has changed is the type names of many enumerations. I'm thinking that this is a value proposition in the long term, since it representations better reuse, but I do think it's going to drive a lot of comment
Grahame Grieve (Dec 30 2019 at 00:59):
ok. released and in production for the validator and publisher
James Agnew (Jan 02 2020 at 19:50):
FYI to anyone following along, I'm 90% of the way through getting this new stuff incorporated into HAPI FHIR 4.2.0-SNAPHOT. Will definitely be landed in the next few days.
Grahame Grieve (Jan 02 2020 at 19:55):
Thx. Can we talk about the license? Why are there two license statements?
Jens Villadsen (Jan 02 2020 at 19:55):
does it 'only' deal with core stuff or does it also deal with profiling generation?
Grahame Grieve (Jan 02 2020 at 19:56):
No. Not profile generation yet
James Agnew (Jan 02 2020 at 19:57):
ohhh..... didn't even notice the two license statements in there. The ASL2 one is most likely being added automatically by a Maven plugin
James Agnew (Jan 02 2020 at 19:57):
(We have one in the HAPI build that automatically adds it to any files that don't have it, that plugin probably got inherited by org.hl7.fhir.core)
Jens Villadsen (Jan 02 2020 at 20:02):
ok ... I guess I'll need to have a look then ... so to understand @James Agnew correctly - ... how does this affect the 'LaF' of the current core classes when updating to 4.2.0? Do you intend to keep it backwards compatible?
James Agnew (Jan 02 2020 at 20:06):
it hasn't been a big deal that I've been able to see.. As Grahame pointed out, the enums have changed a bit so some of my existing code had a few import statements to be fixed, but it was very trivial
James Agnew (Jan 02 2020 at 20:07):
Grahame- I'm assuming you want to stick with MIT license for this code? If so I'll just configure the plugin to insert that instead of ASL.
Jens Villadsen (Jan 02 2020 at 20:08):
I also guess it could come down to the extent of inheritance that one uses on top of the core classes ...
Jens Villadsen (Jan 02 2020 at 20:10):
the doc states its only for R5 - but since the HAPI framework is built a bit agnostic around the FHIR version, I guess this will impact on earlier releases as well?
James Agnew (Jan 02 2020 at 20:15):
Nope, this affects R5 structures only, so no impact to existing applications on earlier versions
Grahame Grieve (Jan 04 2020 at 05:39):
@James Agnew yes let's simply the license to just the MIT one
Last updated: Apr 12 2022 at 19:14 UTC
