Tango vs Phobos

[Koroskin Denis]

On Mon, 11 Aug 2008 20:03:29 +0400, Koroskin Denis <2korden at gmail.com>  
wrote:

> On Sun, 10 Aug 2008 13:18:20 +0400, Lars Ivar Igesund  
> <larsivar at igesund.net> wrote:
>
>> Nick B wrote:
>>
>>> Sean
>>>
>>> So what is the plan to transition Tango to D2.0 ?
>>>
>>> Is the Tango team waiting for const to be removed from D2.0, or will
>>> Tango continue to use D1.0 forever, or is the team waiting to see the
>>> final form of D2.0 before deciding what to do ?
>>
>> We don't expect const to be removed :) As it is, this must be resolved
>> before it is possible to properly port Tango:
>>
>> http://d.puremagic.com/issues/show_bug.cgi?id=2267
>>
>> 1644 was fixed now, but is in our opinion the lesser solution to the
>> problem - we'd much rather have 1961.
>>
>> 2204 is still open I think.
>>
>> Also as long as closures are allocated on stack, that is likely to be  
>> rather
>> detrimental to the performance.
>>
>> Once we have a resolution on that, work on the Tango D 2.0 branch will
>> probably continue, and this branch will be available to all who need  
>> it. If
>> possible, an official release may happen shortly after the time Walter
>> calls D 2.0 to be stable.
>>
>> One other concern is that it is almost impossible to have code that is  
>> both
>> D1 and D2 compatible, something which mean the mantainance of two  
>> branches,
>> a potentially daunting task - it would be good if the "syntactical
>> correctness" restriction on versioned out blocks could be removed for at
>> least D1/D2 identifiers.
>>
>
> We need some other kind of `version' for this, I'm afraid. Something  
> like this, maybe:
>
> #version (StructInsteadOfClass)
>     struct Foo {
> #else
>     class Foo {
> #end
>         void bar() {}
>     }

I would also like to have something like this:

#if CompileTimeConfig.TargetOS == OS.MacOS &&  
CompileTimeConfig.TargetOSVersion >= 10_3_2
    // do something
#else
    // do something else
#endif

Note that it is exactly like static if (sameCondition) { } (i.e. it can  
benefit from CTFE etc), but allows wrong code in false clause.