Expose ValidationErrors as IEnumerable to Prevent Side Effects

[KyleMcMaster]

• Reverts closes #161 Auto-Evaluation of the Status property when Errors ou ValidationErrors are added to the result. #162
• Fixes Change the Status to Failed when any ValidationError is added to the Result. #161 based on discussion and feedback

Sample unit test showing behavior

[Fact]
public void DoesNotAddValidationErrorToValidationErrors()
{
    var result = Result.Invalid(new List<ValidationError>());

    var errors = result.ValidationErrors.ToList();
    errors.Add(new ValidationError());

    result.Status.Should().Be(ResultStatus.Invalid);
    result.ValidationErrors.Should().HaveCount(0);
    errors.Should().HaveCount(1);
}

[github-actions]

Package	Line Rate	Branch Rate	Complexity	Health
Ardalis.Result.Sample.Core	18%	18%	63	❌
Ardalis.Result.FluentValidation	86%	50%	6	✔
Ardalis.Result	21%	0%	82	❌
Summary	23% (95 / 416)	17% (9 / 54)	151	❌

[ardalis]

Should we have more options for including correlation ids? And if so, rather than having separate methods for all of them, would it make sense to bake it into a single parameter object, like:

public record ErrorList(IEnumerable<string> ErrorMessages, string? CorrelationId);

?

Then the factory methods would just take in (ErrorList errorList) and we wouldn't need any overloads or alternate methods.

[KyleMcMaster]

Should we have more options for including correlation ids? And if so, rather than having separate methods for all of them, would it make sense to bake it into a single parameter object, like:

public record ErrorList(IEnumerable<string> ErrorMessages, string? CorrelationId);

?

Then the factory methods would just take in (ErrorList errorList) and we wouldn't need any overloads or alternate methods.

I agree we should try to pass CorrelationId whenever we can, I'll update the methods in a separate PR.