I was wondering if the operation implementation supports sending an
operation and then destroying the operation handle without checking the
result.
Same question through CORBA.
Would that work ?
Sylvain
Orocos Real-Time ToolkitSmarter realtime. Safer threads
|
|
Fire-and-forget operations
I was wondering if the operation implementation supports sending an Same question through CORBA. Would that work ? Sylvain |
Fire-and-forget operations
On Wednesday 17 November 2010 17:24:39 Sylvain Joyeux wrote:
> I was wondering if the operation implementation supports sending an
> operation and then destroying the operation handle without checking the
> result.
>
> Same question through CORBA.
>
> Would that work ?
Yes. We have been discussing/fixing this a while ago. You're welcome to test as
well. Note that a send() through corba does go over the network, and delivers
the 'payload'. So a send() can still block in case there are network problems
etc.
Peter
Fire-and-forget operations
On 11/17/2010 05:39 PM, Peter Soetens wrote:
> On Wednesday 17 November 2010 17:24:39 Sylvain Joyeux wrote:
>> I was wondering if the operation implementation supports sending an
>> operation and then destroying the operation handle without checking the
>> result.
>>
>> Same question through CORBA.
>>
>> Would that work ?
>
> Yes. We have been discussing/fixing this a while ago. You're welcome to test as
> well. Note that a send() through corba does go over the network, and delivers
> the 'payload'. So a send() can still block in case there are network problems
> etc.
Good point, thanks for reminding me.
We only use operations on reliable connections (same machine or
dedicated ethernet connection), so it won't have a big effect.