MS extend C++ significantly for Windows8... and Andrei got name drop

[Manu]

On 26 January 2012 00:55, Adam Wilson <flyboynw at gmail.com> wrote:

> On Wed, 25 Jan 2012 14:44:09 -0800, Manu <turkeyman at gmail.com> wrote:
>
>  On 26 January 2012 00:37, Adam Wilson <flyboynw at gmail.com> wrote:
>>
>>  On Wed, 25 Jan 2012 14:28:46 -0800, Manu <turkeyman at gmail.com> wrote:
>>>
>>>  On 25 January 2012 23:59, Adam Wilson <flyboynw at gmail.com> wrote:
>>>
>>>>
>>>>  On Wed, 25 Jan 2012 13:35:38 -0800, Manu <turkeyman at gmail.com> wrote:
>>>>
>>>>>
>>>>>  On 25 January 2012 21:47, bls <bizprac at orange.fr> wrote:
>>>>>
>>>>>
>>>>>>  On 01/25/2012 07:03 AM, Manu wrote:
>>>>>>
>>>>>>
>>>>>>>  This is fairly interesting. MS have extended their C++ compiler
>>>>>>>
>>>>>>>  significantly for Windows8 with a bunch of non-standard stuff.
>>>>>>>> FINALLY implement garbage collection, ref counting, properties,
>>>>>>>> delegates, events, generics, etc...
>>>>>>>> If other compilers adopt this tech, D loses some advantages.
>>>>>>>>
>>>>>>>>
>>>>>>>>  But you still have to fight with ifndef ,forward declaration, and a
>>>>>>>>
>>>>>>> template syntax against common sense.  Even if you paint shit yellow
>>>>>>> it's
>>>>>>> not necessarily gold.
>>>>>>>
>>>>>>>
>>>>>>>  True, but I think this will mitigate a lot of the motivation Windows
>>>>>>>
>>>>>> devs
>>>>>> have to seek another language if they're not developing cross platform
>>>>>> apps.
>>>>>>
>>>>>> Sadly, since WinRT requires using these language extensions to
>>>>>> interface
>>>>>> with the new windows runtime, you won't be able to write a Windows8
>>>>>> app
>>>>>> in
>>>>>> D.
>>>>>> Interestingly though, D supports almost everything they've added to
>>>>>> C++. I
>>>>>> wonder if it would be possible to do extern(Windows8) to produce a
>>>>>> compatible ABI for linking with MS C++ apps?
>>>>>>
>>>>>> The most interesting features are 'ref new' and 'gcnew', which makes
>>>>>> me
>>>>>> wonder, since Windows8 has an OS garbage collector, would it be at all
>>>>>> possible to have D use the Windows8 GC? I'd prefer this to using D's
>>>>>> own
>>>>>> GC
>>>>>> if it would be supported, and obviously this would be a requirement
>>>>>> if D
>>>>>> was going to interact with WinRT properly.
>>>>>> Also, WinRT uses 'ref new' to allocate ref counted (effectively COM to
>>>>>> my
>>>>>> understanding) objects. I think I read somewhere that D already has
>>>>>> extern(COM) no? I wonder if Windows8 ref type linkage is already
>>>>>> technically supported in D?
>>>>>>
>>>>>>
>>>>>>  There is no Win8GC, it's all ref counted. WinRT is COM with extras
>>>>> and as
>>>>> such should be accessible to D. It would need some extra glue code over
>>>>> what we have now ... like the IInspectable interface.
>>>>>
>>>>>
>>>>>  Really? So what's 'gcnew' for?
>>>>
>>>>
>>> That's for targeting the CLR (.NET) so it doesn't conflict with
>>> new/delete
>>> in regular C++. It goes all the way back to the first C++/CLI in Visual
>>> Studio 2005.
>>> http://msdn.microsoft.com/en-****us/library/te3ecsc8.aspx<http://msdn.microsoft.com/en-**us/library/te3ecsc8.aspx>
>>> <http**://msdn.microsoft.com/en-us/**library/te3ecsc8.aspx<http://msdn.microsoft.com/en-us/library/te3ecsc8.aspx>
>>> >
>>>
>>
>>
>> So there is a GC... It just happens to be the .net GC. Is that a problem?
>> Obviously it's accessible in C++ code. Can you use it to allocate C++
>> objects, or is it exclusively for some sort of interaction with .net?
>> If the rest of the platform is using it...
>>
>
> It's exclusively for the .NET Framework and requires the .NET Framework to
> be installed and loaded to run. It's the .NET Runtime, and as such, it
> would be VERY unwise to base anything D on it.


.NET will be a requirement on all Win8 machines. You didn't answer though,
can you use the .NET GC to allocate C++ objects to use throughout your
native program? Or is gcnew strictly for interfacing C++ with .NET, like
C++/CLI?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/digitalmars-d/attachments/20120126/c5a26261/attachment.html>