Working with a generic class that uses a type that should be of generic type

I have followed Nick Chapsas' tutorial to avoid throwing a ValidationException to treat validation errors and instead rely on LanguageExt.Common.Result<> from LanguageExt library.

I have managed to develop a working solution that relies on MediatR (which uses queries, commands and validators for those), but I feel I am violating the DRY principle.

The relevant code is below:

A sample command

public class CreateDummyModelCommand : IRequest<Result<int>>
{
	public string? Code { get; set; }
	public string? Name { get; set; }
}

/// <summary>
/// this relies on Nick Chapsas idea of not using ValidationExceptions, but return Values:
/// https://www.youtube.com/watch?v=a1ye9eGTB98
/// </summary>
internal class CreateDummyModelCommandHandler : IRequestHandler<CreateDummyModelCommand, Result<int>>
{
	private readonly IValidationUtils<CreateDummyModelCommand, int> _validationUtils;

	public CreateDummyModelCommandHandler(IValidationUtils<CreateDummyModelCommand, int> validationUtils)
	{
		_validationUtils = validationUtils;
	}

	public async Task<Result<int>> Handle(CreateDummyModelCommand request, CancellationToken token)
	{
		var validationResult = await _validationUtils.IsValid(request, token);
		if (validationResult.IsFaulted)
			return validationResult;

		var r = new Random();
		return await Task.FromResult(r.Next(1000));
	}
}

Result<A> provides an implicit cast operator that allow to return a Result<A> and the function to return an A.

I will add ValidationUtils class for completeness, but I think it is less relevant in this case:

internal class ValidationUtils<TReq, TRes> : IValidationUtils<TReq, TRes> 
	where TReq : class, IBaseRequest
{
	private readonly IValidator<TReq> _validator;

	public ValidationUtils(IValidator<TReq> validator)
	{
		_validator = validator;
	}

	public async Task<Result<TRes>> IsValid(TReq request, CancellationToken token = default)
	{
		var validationResult = await _validator.ValidateAsync(request, token);
		if (!validationResult.IsValid)
		{
			var validationException = new ValidationException(validationResult.Errors);
			return new Result<TRes>(validationException);
		}

		var dummyObj = Activator.CreateInstance<TRes>();
		return dummyObj;
	}
}

What I do not like is that virtually any command must include that validation logic at the beginning and I feel that it can be put in a MediatR pipeline behaviour that would execute for each command and trigger the validation, if one is defined. My non-working code is currently the following:

internal class ValidationBehaviour<TRequest, TResponse> : IPipelineBehavior<TRequest, TResponse>
	where TRequest : class, IRequest<TResponse>, IBaseRequest
	where TResponse: struct
{
	private readonly IValidationUtils<TRequest, TResponse> _validationUtils;

	///
	public ValidationBehaviour(IValidationUtils<TRequest, TResponse> validationUtils)
	{
		_validationUtils = validationUtils;
	}

	public async Task<TResponse> Handle(TRequest request, CancellationToken token,
		RequestHandlerDelegate<TResponse> next)
	{
		bool isResult = typeof(TResponse).GetGenericTypeDefinition() == typeof(Result<>);
		if (!isResult)
			return await next();

		var validationResult = await _validationUtils.IsValid(request, token);
		if (validationResult.IsFaulted)
			return validationResult as dynamic;

		return await next();
	}
}

This compiles, but it does not work because since TResponse is actually a Result<>, validationResult is a Result<Result<>>.

One way would have been to work with another generic type to indicate the underlying type of the Result<>, but I cannot change the behavior's signature. While using Result<> usage is convenient, I could break this dependency and use my own class, but I feel I am reinventing the wheel here.

Any idea about how to make this work?

posted over 2 years ago

CC BY-SA 4.0

Alexei‭

5003 reputation 112 101 690 486