Nullable Reference Types analysis can be thrown off by condition

[danwalmsley]

See screenshot here: Iv enabled NRT im on .net core 3.0 release and latest visual studio.

then if I add in a call with a condition on a value being not null...

It seems to suggest to the compiler that after this line, x might be null, because the programmer checked for it being null.

Is this intended behavior?

My helper methods are defined as:

[MethodImpl(MethodImplOptions.AggressiveInlining)]
        [ContractAnnotation("condition:false=>stop")]        
        public static void Requires<TException>(bool condition) where TException : Exception, new()
        {
            if (!condition)
            {
                throw new TException();
            }
        }

[333fred]

Contract.Requires<TException> needs a [DoesNotReturnIf(false)] on the condition. See how Debug.Assert does it. If that still doesn't fix it, I'd suggest opening an issue on dotnet/roslyn, as this isn't the right venue.

[333fred]

Also, yes, the warning is intended behavior. != null is a pure null test: https://github.com/dotnet/csharplang/blob/c229cae634/meetings/2019/LDM-2019-02-20.md#deliberate-tests-for-nullability

[danwalmsley]

@333fred many thanks for feedback, I will test your suggestions.