Checking if NDArray is null returns wrong results

[Superpiffer]

If I try to check if an NDArray is null, I get strange behavior:

NDArray n = null;
bool b1 = n is null;  // TRUE
bool b2 = n == null;  // FALSE!

I saw in NDArray.Operators.cs that if the second value is null, the result is false:

TensorFlow.NET/src/TensorFlowNET.Core/NumPy/NDArray.Operators.cs

Lines 28 to 29 in ce3ddb2

	public static NDArray operator ==(NDArray lhs, NDArray rhs)
	=> rhs is null ? Scalar(false) : new NDArray(math_ops.equal(lhs, rhs));

Is that a wanted behavior? I think that if both values are null, the result should be true.

[m0nzderr]

Just got into this after migrating from the older version that used NumSharp. Almost had a heart attack by watching an expression if (x != null) throwing a NullReferenceException.
It looks like it was an intended behavior, according to the implementation, but wait, this is not a Python anymore! Despite there is an "is" operator in C# now, this is not a common idiomatic form. No one should expect from "x == null" anything else except checking for null! In other words, let's not try to turn C# into a Python, please?

[Superpiffer]

Solved 5821275