Re: [certi-dev] Requesting object attribute values for attributes that no-one has ever owned

[Timi Tuohenmaa]

2012/3/2 Eric Noulard <address@hidden>:
> 2012/3/2 Timi Tuohenmaa <address@hidden>:
>> Hi,
>>
>> It looks to me that sending requestObjectAttributeValueUpdate() for
>> multiple attributes where some of the attributes are owned by no-one
>> (and has never been owned) it sends request to owner of the object
>> even though owner has not published such an attribute.
>
> This looks like a bug, could you file a bug report for that
> https://savannah.nongnu.org/bugs/?group=certi

I made some sort of short bug report of it:
https://savannah.nongnu.org/bugs/?35719

>> In my understanding requesting such unowned attributes should go to
>> bit heaven and those attributes that are owned by someone will get the
>> requests.
>
> Likewise requesting an unpublished attribute should not do trigger
> an update for this attribute.
>
>> I admit that I haven't made too good testing about this
>> issue so if you say this should work, I'll make real tests.
>
> No I think it may not have been thoroughly tested by us.
>
>> However  using Portico the same code works without sending
>> requests for unowned attributes.
>
> +1 for Portico then.
>
>> For the most parts Portico and CERTI seems to works the same (as
>> should of course) :)
>
> No we are definitely better
> Just kidding of course :-]
>
> Can I ask why you tried CERTI if you used Portico before?
> Or may be you are doing some comparison between RTI?

Portico has quite a few small bugs too. I have been able to go around
them, but it haven't been updated at all for a long time so I wanted
to check other options.

Also I have been doing my master's thesis about HLA and some sort of
real-time simulation and newest papers I have found involves CERTI
(and some are written partly by you), so CERTI does not seem to be
dead in any way. So thank you for writing some up-to-date articles
that I can refer :)


Kind regards,
Timi Tuohenmaa