Unsafe accessors can be used to access private members of a class, just like you would with reflection. And the same can be said about changing the value of a readonly field. Let’s assume the following class: Let’s say that for some reason you want to change the value of that read-only field. You could…
In a previous article we covered how to access private members using UnsafeAccessor. This time around, we want to look at it’s performance compared to Reflection, and to see whether it’s truly zero-overhead or not. We’re going to do four benchmarks. If you want to run the benchmarks yourself, you have the code below: Benchmark…
Reflection allows you to obtain type information at runtime and to access private members of a class using that information. This can be particularly useful when dealing with classes outside of your control – provided by a third-party package. While powerful, reflection is also very slow, which is one of the main deterrents in using…
Starting with C# 12, a new ExperimentalAttribute is introduced allowing you to mark types, methods, properties or assemblies as being experimental features. This will trigger a compiler warning during usage which can be disabled using a #pragma tag. The Experimental attribute requires a diagnosticId parameter to be passed in the constructor. That diagnostic ID will…
The ref readonly modifier enables a more transparent way of passing read-only references to a method. Passing of readonly references was already possible in C# by using the in modifier ever since version 7.2, but that syntax had some limitations, or rather too little constraints. So how does the new modifier work? Let’s assume the…
The easiest way to shuffle an array in C# is using Random.Shuffle. This method has been introduced in .NET 8 and works both with arrays and spans. The suffling is done in-place (the existing array/span is modified, as opposed to creating a new one and leaving the source unchanged). In terms of signatures, we’ve got:…
There are some situations in which creating a whole new IJsonTypeInfoResolver will seem overkill, when the default (or any other already defined) type info serializer could do the job with only one or two small modifications. Until now, you could play with the DefaultJsonTypeInfoResolver.Modifiers property for the default type info resolver, but you didn’t have…
A new extension method – GetFromJsonAsAsyncEnumerable<T> – has been added to the HttpClient part of .NET 8. This new mthod will take the response body JSON and deserialize it into an async enumerable operation. The complete signature of the extension method is as follows: Let’s take a look at how to use it. First thing…
Part of .NET 8, JsonNode and JsonArray get a few new additions to their API. We’ve already covered deep copy and deep equality in an earlier article, but there’s more. GetValueKind Returns the JsonValueKind of the current instance. GetPropertyName Returns property name of the current node from the parent object. Throws an InvalidOperationException if the…
Starting with .NET 8, several new methods have been added to the JsonNode class to help with the deep cloning of nodes and checking whether they are equal or not. The DeepClone() method will create and return a deep clone of the current node and all of its descendants. On the other hand, DeepEquals() will…
Starting with .NET 8 you can disable the default reflection-based serializer that comes with System.Text.Json. This can be useful in trimmed and native AOT applications where you don’t want to include the reflection components in your build. You can enable this feature by setting the JsonSerializerIsReflectionEnabledByDefault property to false in your .csproj file. As a…
Starting with .NET 8, the JsonSerializerOptions class features a new TypeInfoResolverChain property in addition to the existing TypeInfoResolver property. With this new property, you are no longer required to specify all the resolvers in the same place, instead, you can add them later as needed. Let’s look at an example: Besides adding new type resolvers…