Re: [certi-dev] ?==?utf-8?q? Ownership Transfer

[CHAUDRON Jean-baptiste]

Hello,

Yes it is a known issue which has been listed and investigated.
Please have a look at:
https://savannah.nongnu.org/bugs/?56284

I have implemented a patch in a dedicated branch:
http://git.savannah.nongnu.org/cgit/certi.git/log/?h=br_bug_56284
Unfortunately, I had not much time to investigate further and test it to fully integrate this to the master branch.
For now, it stays like this.

I hope it helps,
Best regards,
Jean-Baptiste

On Wednesday, May 06, 2020 17:30 CEST, Moritz Gütlein <address@hidden> wrote:
 
> Great, it works! Thank you! I am working with 1516-2010.
> I am curious about another thing: evokeMultipleCallbacks(min,max)
> returns instantly (at least earlier than "min").
> Does the call to enableAsynchronousDelivery switch the whole thing to
> the HLA_IMMEDIATE callback model? Or why is there no waiting for the
> minimum waiting time?
>
> Thanks again,
> Moritz
>
> Am 06.05.20 um 09:17 schrieb Pierre Siron:
> >
> > Le 05/05/2020 à 20:17, Moritz Gütlein a écrit :
> >> Hello everybody,
> >
> > Hello Moritz
> >
> >
> >> I hope this is the right place to ask for help, I didn't find
> >> something like a user-list:
> >
> > This is the good place, wa are using only this list.
> >
> >
> >
> >> So, I am trying to transfer the ownership of an object between two
> >> federates, while updating an attribute. Both federates are running the
> >> following snippet in an infinite loop.
> >>
> >>> while(true){
> >>> if(!currentlyOwning)
> >>> {
> >>>     fed->rtiamb->attributeOwnershipAcquisition(...);
> >>>     while(!fed->rtiamb->isAttributeOwnedByFederate(...)){ >
> >>> evokeMultipleCallbacks(...);
> >>> }                                                           // else{
> >>>     fed->rtiamb->updateAttributeValues(...);
> >>>     while(!shouldDivest){
> >>>         evokeMultipleCallbacks(...);
> >>>     }
> >>>     fed->responseToOwnershipReleaseRequest( amb->releaseObject,
> >>> amb->releaseAttributes );
> >>> }
> >>> advanceTime(...);
> >>> }
> >>
> >> From my understanding this should be alright, but the whole thing
> >> works only if I replace the evokeMultipleCallbacks(...) with
> >> timeAdvance(...) calls. Could you please tell me, if I am doing
> >> something wrong? Is it really necessary to call timeAdvance(...) in
> >> the inner loops?
> >
> >
> > The callback is processed (I have not reviewed the code before to
> > answer) like a RO (Receive Order) message,
> > these messages are (by default) delivered during a time advancing phase.
> > You can change this behavior by invoking the enableAsynchronousDelivery
> > service.
> >
> > Are you working with 1.3 or 1516-2010 ?
> >
> > Bien cordialement,
> > Pierre
> >
> >
> >
> >
> >>
> >> Thank you! Best,
> >> Moritz
> >>
> >
>
> --
> CERTI-Devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/certi-devel