Boost logo

Boost :

Subject: Re: [boost] [Fibers] asio handler not respecting scheduler
From: Eugene Yakubovich (eyakubovich_at_[hidden])
Date: 2014-01-16 16:50:49


On Thu, Jan 16, 2014 at 2:56 PM, Oliver Kowalke
<oliver.kowalke_at_[hidden]> wrote:
> 2014/1/16 Eugene Yakubovich <eyakubovich_at_[hidden]>
>
>
>> Shouldn't the scheduler (run method that
>> follows) decide which fiber to run next? What happens if there's
>> another fiber with a higher priority that's ready?
>>
>
> the scheduler does round-robin, priority is not used

The default scheduler is round-robin but isn't it the point that it
can be replaced with something else that can use priority? What's the
point of having a replaceable scheduler if other parts of the code
resume the fibers themselves?

>
> yield() does suspend the current fiber and puts it at the end of the
> ready-queue, so
> the state of this fiber must be READY

Makes perfect sense. I was not thinking.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk