[dmd-concurrency] real

[Walter Bright]

Enables is I think a better word.

Portability is not the overriding goal of D. Gratuitous non-portability 
is a bad idea, but I think a systems language shouldn't be dumbed down 
to the lowest common denominator. It'd be like crippling 64 bit D so it 
cannot allocate more memory than 32 bit D could. People paid for their 
machine, and they should be able to run apps that fully exploit them.

Also, when compiling for the odd machine that won't support real 
atomics, you get a compile time error, not subtly wrong results.

Andrei Alexandrescu wrote:
> That fosters non-portable code.
>
> Andrei
>
> Walter Bright wrote:
>> Since D is a systems language, it should work where available and not 
>> (compile time error) when not available. Leave it to the programmer 
>> to select the most appropriate workaround where it does not work.
>>
>> Andrei Alexandrescu wrote:
>>> It looks like we're converging on this:
>>>
>>> - shared pointer read/assignment works
>>>
>>> - shared numeric data read/assignment works, EXCEPT for real
>>>
>>> Real is an odd beast - it has no guaranteed width and is mostly 
>>> system-dependent. Should we guarantee atomic assignment to 
>>> shared(real), leave it to the platform, or just disallow it entirely?
>>>
>>>
>>> Andrei
>>> _______________________________________________
>>> dmd-concurrency mailing list
>>> dmd-concurrency at puremagic.com
>>> http://lists.puremagic.com/mailman/listinfo/dmd-concurrency
>>>
>>>
>> _______________________________________________
>> dmd-concurrency mailing list
>> dmd-concurrency at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/dmd-concurrency
> _______________________________________________
> dmd-concurrency mailing list
> dmd-concurrency at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/dmd-concurrency
>
>