bugCERTI - Bugs: bug #45361, TAG and transient message

 
 

bug #45361: TAG and transient message

Submitter:  None
Submitted:  Sat 20 Jun 2015 12:14:35 AM UTC
   
 
Category:  None Severity:  * 3 - Normal
Item Group:  Software error Status:  Fixed
Privacy:  Public Assigned to:  jb_ch
Open/Closed:  Open Release:  * GIT-master
Reproducibility:  Every Time Fixed Release:  NextRelease
* Mandatory Fields

Add a New Comment Rich Markup
   

Mon 18 Dec 2017 04:23:10 PM UTC, comment #3: 

JbCh's Branch dev-scalian/nmprime-impl fixing this bug has been merged into dev-scalian/cvannier at commit #0352af

Please keep us posted if you encounter any more issue.

Clément Vannier <cvannier>
Group Member
Mon 18 Dec 2017 11:23:53 AM UTC, comment #2: 

This behavior has been analyzed and the Null message Prim algorithm has been updated on a dedicated branch.

The new implementation of the Null message Prim algorithm will be integrated in the next planned release.

JbCh <jb_ch>
Group administrator
Mon 06 Jul 2015 06:15:43 PM UTC, comment #1: 

After a discussion on CERTI's devel, the issue seems to be on using NULL Prime messages. When running my application without it, it runs fine.

I attached a simple C++ simulation that exhibits the same behaviour. Unzip and read the  src/README for compiling/using.

(file #34385)

Anonymous
Sat 20 Jun 2015 12:14:35 AM UTC, original submission:  

Hi,

I encountered an odd behavior when using CERTI (lasted git version)  within my Java Framework (the HLA/Ptolemy plugin, see http://ptolemy.eecs.berkeley.edu/ for more info on Ptolemy).
.I'm also using the latest version of JCERTI available on the git repo .

My simulation is a ping-pong between 2 federates, each of them has a lookahead of 0.1 and both start at 0.

There are 2 objects in the federation, drone and calcul.
A publishes and updates drone and has subscribted to calcul, and vice-versa for B.

Here a quick ascii-art sequence diagram :

A --0.1-->Drone--> B
|  <--Calcul--0.2--|
|  --0.3-->Drone-->|
 

Both use the NER time advance mechanism and ask NER(100) at each time.

The sequence stops after the first ping-pong exchange because the RTI grants to B access to time 100.

This bothers me because there is a transient UAV message timestamped at 0.30000000000000004 from A to B but B never receives it before the TAG.

Please find attached the full logs from A and B.

I don't know  if :
- There is a bug in CERTI
- There is a bug in JCERTI
- There is a bug in my code

Any help is welcomed.

Thanks.

PS: Also posted on certi-devel, don't know which one is the most suited.

Anonymous

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

Attach Files:
   
   
Comment:
   

Attached Files
file #34385:  source.zip added by None (25KiB - application/zip)
file #34271:  A_good added by None (14KiB - application/octet-stream - Logs were mixed up thanks to threading. Good version now (with B's lookahead = 0.1))
file #34272:  B_good added by None (13KiB - application/octet-stream - Logs were mixed up thanks to threading. Good version now (with B's lookahead = 0.1))
file #34269:  A added by None (35KiB - application/octet-stream)
file #34270:  B added by None (37KiB - application/octet-stream)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by cvannier (Posted a comment)
  • -email is unavailable- added by jb_ch (Posted a comment)
  •  

    Follow 8 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2017-12-18 cvannier Fixed ReleaseUnknown NextRelease
    2017-12-18 jb_ch StatusNone Fixed
        Assigned toNone jb_ch
    2015-07-06 None Attached File- Added source.zip, #34385
    2015-06-20 None Attached File- Added A_good, #34271
        Attached File- Added B_good, #34272
    2015-06-20 None Attached File- Added A, #34269
        Attached File- Added B, #34270

    Back to the top

    Powered by Savane 3.13-cf05.
    Corresponding source code