Subject: Re: [boost] Boost.Fiber review January 6-15
From: Nat Goodspeed (nat_at_[hidden])
Date: 2014-01-10 09:34:11
On Jan 10, 2014, at 5:20 AM, Oliver Kowalke <oliver.kowalke_at_[hidden]> wrote:
> 2014/1/10 Paul A. Bristow <pbristow_at_[hidden]>
>> And document the schedulers in a separate 'Implementation' section.
May I suggest "library extension" or "customization" instead? I think it's important to allow user-supplied schedulers - e.g. the complaint about the containers used by the default scheduler implementation. Such users could, if desired, supply a scheduler with containers more to their liking.
(I must admit I was a bit floored by the remark about "slower than threads," since that completely disregards the cost of preemptive kernel context switching.)
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk