-
Notifications
You must be signed in to change notification settings - Fork 339
20190121 Community Meeting Minutes
Time: 10:00-11:00 UTC (11:00-12:00 CET, 05:00-06:00 EST, 21:00-22:00 AEDT) Location: Join Zoom Meeting https://zoom.us/j/573322038
Kenneth did his task, others no progress - San Antonio and the R4 release determined the priorities
We have a functional alpha, including validation and snapshot generation. There are some rough edges, mainly the new way choice types can be constrained and some new FhirPath-normative functionality. These are not commonly used however, so we feel confident that the majority of users can start working with this release.
It's still only available on our alpha feed, but we will promote it to beta next Monday on release day.
We have decided that FhirPath will be maintained from within the DSTU2 branch too. After fixing all the unit-tests it became clear that we just need to change some of the tests to work with DSTU2 instance data, with the benefit that all common code (like FhirPath) will be part of DSTU2, and can "blindly" be copied forwards to the other branches.
We currently have the current 1.x releases ready:
- 1.0 - poco-less parsers, "stricter parsing"
- 1.1 - bug fixes
- 1.2 - R4, PR, "George's bug"
Next monday, we will introduce 1.2-beta (with the first R4 bits) and 1.1-final. For this to happen, we still need to do some preparation:
- Fix R4-bug reported by Brian for the validation of "references" (PR#829) in the R4-alpha
- Hotfix for 1.1: the meta-null-reference-bug (PR#818)
- Hotfix for 1.1: the poco "Permissiveparsing" is missing (I think already present in 1.2)
- This means we will have a 1.1.1 which goes to final monday.
For the 1.2 release, we have to figure out in the next 6 weeks whether we need finer-grained control over the "syntactic" validations introduced by 1.2 (switch some of them off or on separately instead of just having PermissiveParsing) + check performance.
Ewout will start working on the mapping engine for .NET. In the public API, we will probably see a new ElementTree (implementing ITypedElement) which can be used to construct in-memory data that is typed. Also useful for unit-tests. This serves as a target for the mapping engine. Whether the mapping engine is going to be open source has not yet been decided on.
We decide to alternate between 20:00 CET (2pm EST) and the current timeslot to facilitate US participants.
George has started working on moving this kind of code into separate projects, and then later we will move it into a separate recognizable directory "common/". Maybe we will then split that directory off in its own Git Submodule, but a quick Google scan has revealed that submodules are not trivial - will have to see whether that's a good idea.