This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?

Yes.

Andrei

On 06/07/2010 03:25 PM, Sean Kelly wrote:
> This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Done.  The module was never documented, so we can probably avoid even mentioning the removal in the changelog.

On Jun 7, 2010, at 1:30 PM, Andrei Alexandrescu wrote:

> Yes.
> 
> Andrei
> 
> On 06/07/2010 03:25 PM, Sean Kelly wrote:
>> This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?
>> _______________________________________________
>> phobos mailing list
>> phobos at puremagic.com
>> http://lists.puremagic.com/mailman/listinfo/phobos

Sean Kelly wrote:
> This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?
>
> 

I'm not seeing where it conflicts. Concurrency is about message passing, right? I see std.atomics as an analog for C++0x atomics, which gives very low level operations for memory fences and such.

On Jun 7, 2010, at 1:38 PM, Walter Bright wrote:
> 
> Sean Kelly wrote:
>> This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?
>> 
>> 
> 
> I'm not seeing where it conflicts. Concurrency is about message passing, right? I see std.atomics as an analog for C++0x atomics, which gives very low level operations for memory fences and such.

I added all the TDPL atomics stuff in core.atomic.  I could have put it in std.atomics instead I suppose, but that would have prevented its use in druntime and it seems potentially too low-level for phobos anyway.  A bit like how std.concurrency uses spawn() as a user-friendly wrapper around core.thread.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puremagic.com/pipermail/phobos/attachments/20100607/127bb6e8/attachment.html>

Sean Kelly wrote:
> On Jun 7, 2010, at 1:38 PM, Walter Bright wrote:
>>
>> Sean Kelly wrote:
>>> This module is only partially implemented and seems to conflict with using std.concurrency for concurrency stuff.  Should it be removed?
>>>
>>> 
>>
>> I'm not seeing where it conflicts. Concurrency is about message passing, right? I see std.atomics as an analog for C++0x atomics, which gives very low level operations for memory fences and such.
>
> I added all the TDPL atomics stuff in core.atomic.  I could have put
> it in std.atomics instead I suppose, but that would have prevented its
> use in druntime and it seems potentially too low-level for phobos
> anyway.  A bit like how std.concurrency uses spawn() as a
> user-friendly wrapper around core.thread.
> 

Sounds right to me. Dump std.atomics.