Re: A little doubt of methodology

Quoting Leopold Palomo-Avellaneda <[email protected]>:

> A Dilluns 21 Maig 2007 14:26, Peter Soetens va escriure:
>> On Monday 21 May 2007 11:32:37 Leopold Palomo-Avellaneda wrote:
>> > Dear people,
>> >
>> > I have a little doubt. I'm developing a program that use the network. So,
>> > the idea is to have a task that listen the socket. When I create the
>> > task, this task is a periodic task, and an x miliseconds a run the task,
>> > or a nonpeeriodic task, and when the socket receive some data, then this
>> > emits something through the data port?
>>
>> Even if your data packets come in periodically, it won't work well using a
>> periodic task, as the sender will not have the exact period as the
>> receiving task (due to clock drift). Hence, your receiver is better non
>> periodic and blocks on the socket for new data.
>
> Normal ....
>
>> In RTT 1.2.0, there is better support for these cases. The discussion was
>> held here: <https://www.fmtc.be/orocos-bugzilla/show_bug.cgi?id=353>,
>> you'll need to use the discussed trigger() function as well.
>
> Ok,
>
> so reading the thread I understand that I have to create a nonPeriodicTask
> (yes, bad name, but ...) and then the in the update function put the recv
> command of the networks, no?

Exactly.

>
>
> void update()
> {
>    // ... here comes my algorithm...
>    //....
>
>   // ex.	ret = rt_dev_recv(sockfd, msg, sizeof(msg), 0);
>
>    // if some condition is met:
>    this->getActivity()->trigger();
> }

If you write your code like this, start() your component and then wish 
to stop() it again, stop() will wait until update() returns, thus make 
sure that rt_dev_recv has a time-out or so, otherwise, stop() will 
appear to freeze (until a network packet arrives)

Peter

-- 
www.fmtc.be


----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

-- 
Orocos mailing list
[email protected]
http://lists.mech.kuleuven.be/mailman/listinfo/orocos