(
Jump to the original submission)
Mon 29 Mar 2010 06:07:02 PM UTC, comment #6:
Fix works. Thank you.
|
Mon 29 Mar 2010 05:52:24 PM UTC, comment #5:
OK now with a clean stack trace
I found a better place to look :-)
i.e.
certi::ObjectClassBroadcastList::createReducedMessageWithValue
I think I did found the off-by-one array access problem.
I've just checked in a possible fix.
Would update CERTI CVS HEAD and retry?
The trouble I see here is that the current HLA_TestSuite
did not catch that one...
We should definitely have a better non-regression suite.
|
Mon 29 Mar 2010 03:38:19 PM UTC, comment #4:
Hello.
Yes I am using recompiled versions and it does not work even with java port of monitor without PyHLA. I have checked everything and recompiled FlightGear again just to be sure but the results are the same.
Just for record here is the call stack
http://pastebin.com/dAtEKiC1
|
Mon 29 Mar 2010 02:23:13 PM UTC, comment #3:
Ok then let's say I'll wait for you on that one.
|
Mon 29 Mar 2010 02:21:35 PM UTC, comment #2:
I can certainly help here, but not before weekend.
|
Mon 29 Mar 2010 01:54:41 PM UTC, comment #1:
Hi Andrej,
Browsing the C++ code I did not manage to find
how we can generate such invalid access in send interaction.
Before going on with bug tracking could you tell me
if you did recompile PyHLA with the current CERTI CVS?
Same question for FlightGear plugins?
I want to rule out a version mismatch between the libraries.
Normally it shouldn't be possible any more because
of the new OPEN_CONNEXION message but who knows...
Moreover I do not see any "Send Interaction" in monitor-asn.py
so I do not see how they can be sent but in FlightGear plugins.
To go on with strangeness RTIG does not call
setParameters on NM_Send_Interaction at all.
RTIA does.
I'm afraid I won't able to reproduce the failing test case
using flightgear + monitor-asn.py
may be Petr would be able to do it more efficiently than I can.
|
Mon 29 Mar 2010 12:18:22 PM UTC, original submission:
Description:
When trying to run simulation RTIG dies.
Debug shows that it is "vector subscript out of range" located at
NM_Classes.hh / NM_Send_Interaction / setParameters function
(maybe accessing index out of bounds?)
How to reproduce:
I am able to reproduce this bug with VirtualAir simulation setup. My configuration contains
- certi 29. march. '10 CVS Head
- patched flightgear
- monitor-asn.py
After I run RTIG and FlighGear everything works fine but after runnign monitor RTIG crashes.
|
(Note: upload size limit is set to 16384 kB, after insertion of the required
escape characters.)
Attach File(s):
Comment:
No files currently attached
Depends on the following items: None found
Items that depend on this one: None found
Follow 8 latest changes.