On 2009-01-02 20:44:21 -0500, "Jarrett Billingsley" <jarrett.billingsley@gmail.com> said:

> On Fri, Jan 2, 2009 at 8:35 PM, Michel Fortin <michel.fortin@michelf.com> wrote:
>> 
>> How's this different from
>> 
>> case a*.b:
>> 
>> is this:
>> 
>> case ((a*).b):
>> 
>> or is it:
>> 
>> case ((a) * (.b)):
> 
> What the heck is ((a*).b)?  * is not a postfix operator.

It is, kind of: if "a" is a type, postfix "*" makes it a pointer.

If you test it, it'll compile if "a" is a struct having a member "b".

-- 
Michel Fortin
michel.fortin@michelf.com
http://michelf.com/

Michel Fortin wrote:
> On 2009-01-02 21:39:52 -0500, Benji Smith <dlanguage@benjismith.net> said:
> 
>> Think of it like this:
>>
>>    MyClass?.myProperty
>>
>> It's a static field of the nullable MyClass type.
> 
> Just like:
> 
>     MyStruct*.myProperty
> 
> It's a static field of the struct for which you have a pointer to. Exact that for it to compile, you need to write:
> 
>     (MyStruct*).myProperty
> 
> I propose we do things the same way for "?".

That's encouraging. I wasn't able to find any other ambiguous situations (and the one I did find was very obscure); and it does seem (as you pointed out) that ambiguities with a*b are far more common.
It would be an awesome feature IMHO. Null class references are by far the most common bug I make in D. And despite what people have said, working out why a class is null is not always easy, especially when it only shows up rarely in production code.