Shane Krueger
Shane Krueger
Partially done in #3172
The current stream writer is copied from the matching HttpRequestStreamWriter class; it would probably make sense to copy in the reader class also.
See: - https://github.com/graphql-dotnet/graphql-dotnet/blob/master/src/GraphQL.NewtonsoftJson/HttpResponseStreamWriter.cs
> I propose to postpone this. We should to deal with other (more understandable) PRs first. Fine with me. As can be seen in the performance metrics, I do see...
@IdeaHunter > What obstacles do you have? To test precompiled queries, we can easily change a few lines of code to add a `CompiledNode` member to `Document` which doesn't compile...
This PR splits the execution phase into two: compiling and executing. Common sense says that two passes are going to be slower than one pass, requiring more objects to be...
> Do we need this? I would like to run some more benchmarks after our latest updates. Previously there was a definite performance benefit for large queries. The hard part...
> @Shane32 what conclusion will you have on this PR? I don't have time to work on it at the moment. If the conflicts can be fixed, it's pretty easy...
As these changes affect execution (which has not changed much), rather than parsing/serialization (which has), I do not believe the effect will have changed substantially from my comments and performance...
```cs [JsonConverter(typeof(InputsJsonConverter))] public Inputs Variables { get; set; } ```