D for Game Development

[via Digitalmars-d]

On Monday, 10 August 2015 at 19:34:26 UTC, rsw0x wrote:
> On Monday, 10 August 2015 at 19:31:55 UTC, David Gileadi wrote:
>> On 8/10/15 12:25 PM, "Ola Fosheim =?UTF-8?B?R3LDuHN0YWQi?= 
>> <ola.fosheim.grostad+dlang at gmail.com>" wrote:
>>> [...]
>>
>> …[insert your language here] has a long way to go… :)
>
> Which is why I think people are attracted towards D. It's very 
> close to being "there". The large elephant in the room is the 
> garbage collector. Every way to work around it feels like a 
> massive, ugly hack.

Yes, it is not a good fit for D. Although, I find the Pony-lang 
approach interesting, but that is an actor language so it should 
not be compared to D, but to vibe.d.

Pony uses a per heap GC, a cross actor GC, and actor collection 
(collecting the whole actor and heap when the actor cannot 
receive more messages). Each actor (fiber in D) does not have a 
stack, the stack is per thread so I believe you "yield" when the 
stack has been unwound, they only use C-ABI when calling C 
functions.

But they also have an advanced pointer type system that can 
distinguish between at least 6 different reference-aliasing 
situations (or was it 12?).

https://www.youtube.com/watch?v=KvLjy8w1G_U

Anyway, it is refreshing. Maybe D can pick up some ideas from 
Pony.