Ewout Kramer

I now realize this is pretty hard - not because we can't, but because the new validator is version independent (R3/R4 etc), but the code we wrote to validate this...

That's right. If have documented this behaviour here: https://docs.fire.ly/projects/Firely-NET-SDK/parsing/system-text-json-serialization.html#serialization-with-pocos-and-system-text-json. The serializer does not do any validation and can easily be (misused) to generate incorrect json when pushed to do so....

I have been thinking about this a bit. The problem of optimizing for empty objects is that we only know an object is not empty when we hit the first...

We're on thin ice here, since FHIR does not specify a serialization/deserialization for subtrees, as we are discussing here. For example, it's impossible to serialize a `FhirBoolean` to json, because...

Note that Gino is updating his latest codegen so it does no longer produce breaking changes in the generated C# code, so we should try to apply this fix to...

What is it that you want us to do here? It does not seem to be a specific .NET FHIR SDK problem, as you highlight with NewtonSoft. The good news...

This is related to #1221. The 'value' being special is pretty deeply embedded in the code - and since FhirXmlNode has no metadata (so knows nothing about `representation`), there's very...

I did make it work in the MappingLanguage parser (which was based on Superpower). Too bad we cannot use that code anymore. Are we sure we can make this work...

@brianpos - note that both Gino and I had some review remarks. Build is currently failing because of a QA step: https://dev.azure.com/firely/firely-net-sdk/_build/results?buildId=38380&view=logs&j=558a83a0-5585-5f85-5265-cdf5febe3f62&t=a961ae7d-38e2-5a82-9685-674f5235ed2d, we have added new constructors which breaks binary...

> Should I just merge in the whitespace and comments capturing changes into this branch too now? Nah, let's do two PRs, this one is done & reviewed now. Just...