Eliminate "new" for class object creation?

[Robert Jacques]

On Mon, 19 Oct 2009 22:53:23 -0400, Brad Roberts <braddr at puremagic.com>  
wrote:

> Robert Jacques wrote:
>> On Mon, 19 Oct 2009 18:45:01 -0400, dsimcha <dsimcha at yahoo.com> wrote:
>>
>>> == Quote from Andrei Alexandrescu (SeeWebsiteForEmail at erdani.org)'s
>>> article
>>>> I'm having a hard time justifying that you use
>>>> new X(args)
>>>> to create a class object, and
>>>> X(args)
>>>> to create a struct object. I wrote this:
>>>> ============
>>>> The syntactic  difference between  the expression creating  a @struct@
>>>> object---Test(@\meta{args}@)@---and the  expression creating a @class@
>>>> object---\cc{new Test(}\meta{args}@)@---may be  jarring at first. \dee
>>>> could have dropped the @new@  keyword entirely, but that @new@ reminds
>>>> the programmer that an object allocation (i.e., nontrivial work) takes
>>>> place.
>>>> ===============
>>>> I'm unhappy about that explanation because the distinction is indeed
>>>> very weak. The constructor of a struct could also do unbounded amounts
>>>> of work, so what gives?
>>>> I hereby suggest we get rid of new for class object creation. What do
>>>> you guys think?
>>>> Andrei
>>>
>>> Absolutely.  I've thought this for a while but hesitated to bring it
>>> up because I
>>> felt it was a bikeshed issue.  Now that I think of it, though, it
>>> would have the
>>> substantive benefit of making it easier to switch from structs to
>>> classes if you
>>> suddenly realize you need polymorphism, or from classes to structs if
>>> you suddenly
>>> realize you need value semantics.  I really can't see any downside
>>> other than the
>>> loss of static opCall for classes, which doesn't have tons of good use
>>> cases anyhow.
>>
>> As structs mix static opCall and ctors, there's no reason classes can't.
>
> Why do we still have static opcall?  What role does it play that ctors  
> don't?

Well, when you want it to return something other than a new instance. The  
only practical example I have is for singleton-types implemented using  
static (TLS) variables. Also, there was a recent bug/regression with  
regard to struct ctors vs opCall. Though I think it's been fixed.