Null check syntactic sugar; Make foo? equivalent to foo != null

[kingces95]

So here's my weekend idea...

Null checks are common so why not sprinkle on some (more) syntactic sugar?

Instead of

if (myArg == null) throw new ArgumentNullException()

how about

if (!myArg?) throw new ArgumentNullException()

We've already become accustom to an expression that starts with myArg? being declarative syntax for a null check followed by property access (as in myArg?.Prop) or a null check followed by an alternate expression (as in myArg ?? myOtherArg). So if myArg? is followed by nothing, then... well... it's just a null check resulting in a bool (which could possibly be followed by some procedural action like throwing an exception).

[HaloFour]

If obj? produced a bool it would be expected that obj?.ToString() would return either "True" or "False", but that is already legal syntax that has a completely different behavior. This proposal would either break this behavior or introduce ambiguities.

[MillKaDe]

.. has already been suggested a few times .. See:

• #196 - Discussion: null evaluation operator to boolean
• #545 - Proposal: ?obj operator for real '!null' check on all nullable types
• plus several others linked in those two threads ..

Possible syntax (unary prefix operator with highest precedence):

C# (now)	C# (proposed)	C/C++	remarks
if (p != null) ..	if (?p) ..	if (p) ..	= !ReferenceEquals (null, p)
if (p == null) ..	if (!?p) ..	if (!p) ..	= ReferenceEquals (null, p)

The ! operator for bool already exists, and the ? operator for non-null checks would be implemented something like this ..

static public bool operator ? (T t) where T : class => !ReferenceEquals (null, t);

For details about this particular syntax see the two linked threads.

Both threads mentioned above are closed right now, but I still think a null check operator would be really nice.

[MillKaDe]

@HaloFour : I think a prefix instead of postfix operator would avoid the problems you mention.

[HaloFour]

My opinion is that an equality comparison is succinct enough and significantly more readable than either operator-based proposal.

[kingces95]

@HaloFour, ah, yeah. Foo?.ToString() is ambiguous. Duh. Well, that's why language design should be left to the professionals!

@MillKaDe thanks for the links. The prefix ops may work but I gotta say I'm more accustom to seeing the question mark as a postfix operator..