[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [certi-dev] Messages received after resignation
From: |
Arnaud Degroote |
Subject: |
Re: [certi-dev] Messages received after resignation |
Date: |
Wed, 8 Apr 2015 17:20:07 +0200 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
On 08/Apr - 17:04, Eric Noulard wrote:
> 2015-04-08 16:39 GMT+02:00 Arnaud Degroote <address@hidden>
> :
>
> > On 08/Apr - 15:17, Eric Noulard wrote:
> > > 2015-04-08 14:05 GMT+02:00 Arnaud Degroote <
> > address@hidden>
> > > :
> > >
> > > > Hi all,
> > > >
> > > > I'm not an expert in HLA, so I'm wondering if it a normal behaviour, or
> > > > an issue in the RTIG. I observe that I continue to receive some
> > messages
> > > > from RTIG (in particular NM_Reflect_Attribute_Values) after I resign
> > > > from the federation. Is it normal ? Is there something I forget to do ?
> > > >
> > >
> > > No it's not.
> > >
> > > Which version of CERTI are you using ?
> >
> > I use the last release version 3.4.3
> > >
> > > Would you be able to try with git master?
> >
> > The issue is still here in master
> >
>
>
> OK then this is a bug :-)
>
>
> > > I would think that resign should generate "unsubscribe"
> > > but it is possible that no-one hit this scenario with CERTI before
> > because
> > > one usually do
> > >
> > > create and/or join
> > > publish
> > > register
> > > subscribe
> > > ...
> > >
> > > unsubscribe
> > > unregister
> > > unpublish
> > > resign
> > > destroy
> >
> > With a proper call of unsubscribeObjectClass, the message are not
> > emitted anymore.
>
>
> Ok then the bug is still there. We should automagically unsubcribe when
> resigning.
> Would you be kind enough to open a bug report for that?
Will open an issue
>
> Patch proposal are welcomed as well :-)
I will take a look in my spare spare time, so I don't guarantee anything
:).
>
>
>
> > Why is the API not really symetric
> > (subscribeObjectClassAttributes vs unsubscribeObjectClass) ?
>
>
> This is historical assymetry in HLA 1.3
> both
>
> unsubscribeObjectClassAttributes
> and
> unsubscribeObjectClass
>
> exist in the 1516-2000 and 1516-2010 API.
> see <ECRTI-SRC>/include/ieee1516-20*/RTI/RTIambassador.h
>
>
>
> Moreover, I
> > would be more happy with an API to do the contrary of
> > requestObjectAttributeValueUpdate, but I don't find anything to do that.
> > Do I miss something ?
> >
>
> What to you mean by "the contrary" ?
>
> One federate do
> register+UAV (through RTIambassador)
> and some others do
> subscribe (through RTIambassador) and get RAV (through FederateAmbassador)
>
>
> when you do
> ROAV this is an explicit request to get extra value from the publisher even
> if it did not do UAV on his own.
>
> then the owner get PAVU = Provide Attribute Value Update
> and the requester should get back an RAV.
>
> i.e.
> UAV+RAV is a push mechanism between federate
> whereas
> ROAV+PAVU+RAV is a pull mechanism between federate
>
> what other scheme are you thinking of?
>
Nothing, it was an misunderstand on my side, I thought that ROAV is
needed to get update in addition to subscribe. Now, your previous
comment makes sense.
>
>
> > >
> > > so that when resigning we did already unsubscribe.
> > >
> > >
> > >
> > >
> > >
> > >
> > > --
> > > Eric
> >
> > > --
> > > CERTI-Devel mailing list
> > > address@hidden
> > > https://lists.nongnu.org/mailman/listinfo/certi-devel
> >
> >
> > --
> > CERTI-Devel mailing list
> > address@hidden
> > https://lists.nongnu.org/mailman/listinfo/certi-devel
> >
> >
>
>
> --
> Eric
> L'élection n'est pas la démocratie -- http://www.le-message.org
> --
> CERTI-Devel mailing list
> address@hidden
> https://lists.nongnu.org/mailman/listinfo/certi-devel
signature.asc
Description: Digital signature