Add traceId and spanId properties in LogstashJsonFormatter output

[tbdrake]

Fixes #53

See related changes in serilog/serilog#1955

For reference, Serilog.Formatting.Json.JsonFormatter also includes these properties here:
https://github.com/serilog/serilog/blob/cd6420aeafcdedd3790da1a03a419e389f0faa64/src/Serilog/Formatting/Json/JsonFormatter.cs#L74-L84

Also enabled <Nullable> in the unit test project (already enabled in the main project) and fixed some warnings. moved to #56 and #57.

[nojaf]

@Smaug123, I don't see any harm in this, do you?

[tbdrake]

@nojaf One concern I realized is that existing code could be adding traceId and spanId properties already, for example with https://github.com/RehanSaeed/Serilog.Enrichers.Span, and in that case the LogstashJsonFormatter would have duplicate properties. I found this issue #39 so it looks like this applies to other properties like message as well.

I added a failing test in the latest commit and plan to look into #39 and maybe open a separate PR.

[Smaug123]

I don't see any harm in this per se, but I'm certainly not a domain expert, and that issue does look concerning! (On general principles I'd prefer to pull out the Nullable enablement into a separate PR, so that the change with possibly-dangerous semantics can go in on its own.)

[tbdrake]

I don't see any harm in this per se, but I'm certainly not a domain expert, and that issue does look concerning! (On general principles I'd prefer to pull out the Nullable enablement into a separate PR, so that the change with possibly-dangerous semantics can go in on its own.)

Thank you for the input and suggestion! I moved the <Nullable> changes to separate PRs #56 and #57. (Also I accidentally clicked close PR 😄)

[tbdrake]

I added a failing test in the latest commit and plan to look into #39 and maybe open a separate PR.

I adjusted the test to pass based on looking into it further and added comment here on #39.

[tbdrake]

@nojaf No rush but do you have any ETA for a next release, and do you think this change could go in it?

[nojaf]

Thanks, this looks fine. I'll publish a new alpha so you can test this further on your end.