Joshua Marner

Also the majority of the diff issue happens when I split the class in two at the end, and I don't think I can help that because the diff view...

@TysonMN Do you have a set of performance tests or techniques that you used for performance testing?

Test `DynamicViewModel.UpdateModel()` with various numbers of get bindings and consecutive calls to update. Before this PR: | Method | BindingCount | UpdateCount | Mean | Error | StdDev | |-------...

@TysonMN do you want a PR with a benchmarks project? It's currently here: https://github.com/elmish/Elmish.WPF/compare/master...marner2:Elmish.WPF:feature/add_benchmark_project?expand=1

> So the conclusion is that the performance is essentially the same before and after this PR, right? Yes. Seems to be mostly within margin of error, or mixed results....

The PR with the benchmark project is #538

> I don't think your commit message matches the commit. In particular, I don't see any type alias being added. Line 20 is the type alias.

@TysonMN I split the PR into two commits that *actually* match their commit messages now.

You probably need something like this integrated into an FSharp version of the `JsonFormatter` (but for `valueoption`s): ```fsharp type OptionConverter() = inherit JsonConverter() override x.CanConvert(t) = t.IsGenericType && t.GetGenericTypeDefinition() =...

@googleberg current command line command (requires global tool install from nuget): `protoc-gen-fsharp --out=x ...[other args]` Improved version we'd like to write (can be installed locally): `dotnet protoc-gen-fsharp --out=x ...[other args]`...