Am I using this correctly?

[mwasson74]

Hello! Am I using this incorrectly or why does the compiler think that the return of .Null might be null? Is there a way to use it differently to let the compiler know that it won't be null when this returns? Is there a different extension I should be using? Thanks!!

var sptf = Guard.Against.Null(sample.Sptfs.FirstOrDefault(x => x.Key == SptfKey), nameof(SptfKey), "I have no idea why this might be null");
var thisMightBeNull = sptf.Key;

[amal-stack]

@mwasson74 This is happening because the compiler does not have enough information to tell that sptf.Key will not be null after calling Guard.Against.Null(). For now, you can safely suppress this warning using the null forgiving operator (!):

var thisMightBeNull = sptf!.Key

@ardalis This is possibly because Guard.Against.Null is defined as:

public static T Null<T>(this IGuardClause guardClause,
        [NotNull] [ValidatedNotNull] T input,
        [CallerArgumentExpression("input")] string? parameterName = null,
        string? message = null)

If the input parameter definition is changed from:

T input

to

T? input

the warning mentioned in the question goes away.
Since the type T does not have a constraint like struct or class, the T? syntax will only be supported from C# 9. To use the T? syntax prior to C# 9, there must exist a class or struct constraint for the type T. The pre-C# 9 equivalent to T? seems to be the AllowNullAttribute.

Edit: PR #261 which fixes this issue has been merged. The fix is available since version 4.1.0.

[mwasson74]

Thanks, @amal-stack, we ended up using the null-forgiving operator here. I knew I could safely use it but assumed I just wasn't using .Null correctly. Hopefully a small update/PR will allow this to work as I assumed it did.