Boost logo

Boost Users :

From: Hal Vaughan (hal_at_[hidden])
Date: 2008-03-17 11:50:46


Zeljko Vrba wrote:

> On Mon, Mar 17, 2008 at 04:09:53PM +0100, Hughes, James wrote:
>>
>> Or put a usleep(0) in the thread to force the timeslice to expire? IIRC.
>>
> Calling sched_yield() would also probably help, but that's a non-solution
> to
> the OP's non-problem. To the OP: if a particular order of execution is
> desired, it has to be explicitly coded with various synchronization
> primitives.

It's not about order of execution. I explained it in another post, but it
didn't come through right away due to moderation. (Are all posts directly
to the news group moderated? Do I have to post via email for more timely
posting?)

My concern, when I saw one thread execute completely then another is that it
seems one thread hogs the resources, even if it's not doing heavy work.
Since my main task, in the actual program I'm writing, is to monitor
something and report all output on one thread and to have another to allow
commands to go to the device, I want to be sure that the monitoring thread
won't be blocked while data it needs is coming in, but I don't want that
thread to block the other one while it (the monitoring thread) is waiting
for incoming data.

Hal


Boost-users list run by williamkempf at hotmail.com, kalb at libertysoft.com, bjorn.karlsson at readsoft.com, gregod at cs.rpi.edu, wekempf at cox.net