Re: [certi-dev] late arriving federate

[Eric Noulard]

>From the HLA 1.3 document we have

"The Enable Time Regulation service shall enable time regulation for
the federate invoking the service, thereby enabling the federate to
send TSO messages. The federate shall request that its logical time
and lookahead value be set to the values specified as arguments. The
RTI may not be able to set the federate’s logical time to the value
that was requested because doing so might enable the federate to, for
example, send a message with a time stamp smaller than the current
logical time of another federate. The RTI shall indicate the logical
time assigned to the federate through the Time Regulation Enabled †
service. The logical time that is assigned shall be greater than or
equal to that requested by the federate."

and

"Invocation of the Time Regulation Enabled † service shall indicate
that a prior request to enable time regulation has been honored. The
value of this service’s argument shall indicate that the logical time
of the federate has been set to the specified value."

some more precision in 1516 document:
"The logical time provided when time-regulation is enabled shall be
the smallest possible logical time that is greater than or equal to
the joined federate’s current logical time and for which all other
constraints necessary to ensure TSO message delivery are satisfied. In
general, this means that the logical time that the joined federate
will be given (plus the requested lookahead)
must be greater than or equal to the maximum logical time of all
time-constrained joined federates (with the possibility of being of
equal value depending on what form of time advancement was used by
each time-constrained joined federate)."

So Yes CERTI is currently wrong we should build the "Time Regulation Enabled"
built by CERTI should contains

"logical time that the joined federate will be given (plus the
requested lookahead)
must be greater than or equal to the maximum logical time of all
time-constrained joined federates"

I don't know [yet] whether if the patch provided by Mathias is enough or not,
the patch looks good but I did only fast read it.

I don't know whether if I will have much time to look into it more seriously
before next week, anyone wanting to go seriously into this one?

Or may be helping Mathias with the current TimeManagement design.

-- 
Erk
Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org