Revert changes required for ARMv2 experiment

[andyleejordan]

PR Summary

Since we no longer expect DSC to ingest ARMv2 as generated by Bicep (in favor of Bicep invoking DSC resources via the gRPC extension), we can remove this unfinished work that simply got our demos running again.

PR Context

This deserialization approach was a cheap workaround, we had not yet revisited the design to properly support symbolic names (such as fixing resource IDs, dependency resolution, etc.).

We also don't need to add the Bicep emitted fields that we were just skipping.

[andyleejordan]

This is a revert of #1210, but with merge conflicts resolved, and still-relevant tests kept. We can also close #1221 and #1223. I would strongly encourage we remove the copy() (loop) implementation and rely on higher-order tooling for complex resource orchestration.

[Gijsreyn]

@andyleejordan - isn't the copy() functionality a v1 feature?

[andyleejordan]

@andyleejordan - isn't the copy() functionality a v1 feature?

🤷 as far as I could tell it was a lot of hassle for little if any gain as it was really to support for loops that came from Bicep. We don't necessarily need to do that in DSC-native configs, but it's just a suggestion from me and not relevant to this PR (except that I removed the part of the copy() test dealing with symbolic names).

[Gijsreyn]

@andyleejordan - isn't the copy() functionality a v1 feature?

🤷 as far as I could tell it was a lot of hassle for little if any gain as it was really to support for loops that came from Bicep. We don't necessarily need to do that in DSC-native configs, but it's just a suggestion from me and not relevant to this PR (except that I removed the part of the copy() test dealing with symbolic names).

Thanks for the explanation, Andy, and I get that part indeed. Symbolic names are more towards v2 (or even are). There might be other ways to loop for sure, but I can't think of any atm.

[michaeltlombardi]

Code changes look good to me - I think there's still some open questions on impact to other functionality (copy loops), but I prefer that functionality be pushed to higher order tools (bicep, a DSC config builder in PowerShell, whatever) instead of being inherent to the engine / configuration document.