Re: [certi-dev] Trying to use CERTI with regions

[Mattias Helsing]

Hi again 
i got the simplest example down to 90 lines and it uses the Test.fed from the Billard example since that had some Spaces defined. At my end this sample produces RegionNotKnown exceptions

/Mattias

On Fri, Dec 14, 2012 at 2:03 PM, Mattias Helsing <address@hidden> wrote:

Hi,
I whipped up an example to show my problem. I used an example from the test directory and modified it to fit how we use HLA. The tar contains 2 separate programs, where test_RoutingSpacesSimple.cc shows the initial problem we get. That is - calling notifyAboutRegionModified throws RegionNotKnown. The test_RoutingSpaces.cc file is basically identical to test_RoutingSpaces.cc but also creates an object and updates it. This shows some errors that I guess are the product of the previous error . I created a test_Space.fed which I used to reproduce the error here which is included in the tar. I also included the Makefile I used to create the examples for reference. I usually prefer cmake but for tiny programs like this make is easier. I am on Ubuntu linux and using CERT-3.4.1. Anyway it's a start. Tell me what to change and I will.

cheers
Mattias

On Thu, Dec 13, 2012 at 5:36 PM, Mattias Helsing <address@hidden> wrote:

hi again,
and thanks for the quick reply. I'll whip together a small test app as soon as possible.

cheers Mattias

On Thu, Dec 13, 2012 at 2:21 PM, Eric Noulard <address@hidden> wrote:

2012/12/13 Mattias Helsing <address@hidden>:

> Hi all
>
> I wanted to ask if anyone here tried to use CERTI with regions? and in that
> case how did it work out for you?

I'm afraid that region has been unused for quite a time now.
Since we don't currently have a regression test for that as well it has been
untested as well.

so...

> We use CERTI in our simulator setup and running without regions active the
> RTI is working perfectly fine, but as soon as I enable a region I get an
> endless stream of libRTI Exception NetworkError, and UN Socket(EmettreUN) :
> : Broken pipe
>
> From what I can tell I successfully call
>     getRoutingSpaceHandle(..)
>     getDimensionHandle(...)
>     createRegion(...)
> but then fails in
>     notifyAboutRegionModification(*myregion);
> ...which tells me that: Region not known...
>
> I realize this may not be much to go on, but I wanted to describe my problem
> to catch: "Yes, I know what you have done wrong" answers from more
> experienced users. Eventually I'll have to dive into the CERTI code if I
> cant find a solution to this since we really like CERTI.

I would suggest you write a possible standalone test which exhibits your issue
Then I'll help you to:

1) Add this test to our HLA_TestSuite
    http://cvs.savannah.gnu.org/viewvc/applications/HLA_TestsSuite/?root=certi

2) Investigate and solve the issue with you.

> As a note this region stuff has been working fine with DMSO and MÄK, with
> the difference that I have to modify the fed files slightly. So given that I
> have the space MySpace and the dimension MyDim specified in the fed file an
> attribute would be described like this for DMSO/MÄK:
> (attribute MyAttr best_effort receive MySpace)
> ...while with CERTi I have to leave out the MySpace in the above.

I'll check may be this is an error in the CERTI fed parser grammar.

> Sorry about the long post. Any help is greatly appreciated. If further
> information is required I'll gladly get it for you.

Like I said smallest standalone C++ test case would be a very good start.

--
Erk
Le gouvernement représentatif n'est pas la démocratie --
http://www.le-message.org

--
CERTI-Devel mailing list
address@hidden
https://lists.nongnu.org/mailman/listinfo/certi-devel

test_main.cc
Description: Binary data