Re: [Jack-Devel] atomic operations

On Sat, Aug 17, 2013 at 2:33 PM, Kaspar Bumke <[hidden]>wrote:

> Another thought, what about exposing the functions in JackAtomic via the
>> API somehow? It seems like it could be really useful for a lot of jack
>> applications, like jack_ringbuffer is.
>>
>
> What do you think of this? I would be willing to work on a patch if you
> think you might consider it.
>
>
> On 7 August 2013 21:16, Kaspar Bumke <[hidden]> wrote:
>
>> Another thought, what about exposing the functions in JackAtomic via the
>> API somehow? It seems like it could be really useful for a lot of jack
>> applications, like jack_ringbuffer is.
>>
>
I do not recommend JACK exposing this API. Atomic interfaces are hard to
maintain and test across a wide range of processor architectures and OS
platforms.

Ten years ago, I sought other alternatives before introducing some internal
functions needed for certain multi-threading situations, but I could not
find anything suitable.

Now there are better alternatives which are fully maintained, such as the
boost atomic implementation, and the new std::atomic feature in C++11.

 http://www.chaoticmind.net/~hcb/projects/boost.atomic/doc/index.html

JACK should not take the responsibility for maintaining things like that.
<http://www.chaoticmind.net/~hcb/projects/boost.atomic/doc/index.html>--
 joq

1376778414.13583_0.ltw:2,a <CAB6SgyUZmc6+=B8sw+XWd3mBXTLoY32YbUwXeYNomGTeweAtCw at mail dot gmail dot com>