2006/3/9, David Luff <[EMAIL PROTECTED]>:
Melchior is refering to this
#0 0x080c9996 in FGAIPlane::GetLeg (this=0x0) at AIPlane.hxx:80
cptf #1 0x080c1c3f in FGTower::ProcessDownwindReport (this=0xd6bedb8, t=0xddccce8) at tower.cxx:664
cptf #2 0x080c5e95 in FGTower::Respond (this=0xd6bedb8) at tower.cxx:521
cptf #3 0x080c9734 in FGTower::Update (this=0xd6bedb8, dt= 0.033333333333333333) at tower.cxx:357
cptf #4 0x080a5372 in FGATCMgr::update (this=0xb371d90, dt=0.033333333333333333) at ATCmgr.cxx:167
cptf One can really rely on ATC crashing at the right moment. ;-)
(Somehow a plane occurs which contains no AI and is not ourselfs?)
I did the changes you are refering to, because there were a couple of STL problems hidden. I may have broken something so I did some heavy investigations into that and the most plausible cause seems to me that planes with identical callsigns have been around. I introduced some prints, but the problem didn't trigger any more for me. The patch removes some unused code, too.
Melchior FRANZ writes:
> * Curtis L. Olson -- Thursday 09 March 2006 17:59:
> > 2. We need to aggressively hunt down any random crashes
>
> I'm aware of four crashes:
>
> (A) tower.cxx /AI -- old, but very annoying. Happens occasionally.
> Very hard to reproduce. Olaf looked into it and advised me to
> add a debug message in my copy, but since then I had no problems.
> (Maybe I should commit that? ;-) This bug basically makes
> long-distance flights impossible.
>
OK, you've pricked my consciensce. I'll make a concerted effort to track that one down, since it's undoubtably one of mine :-( I thought it might have been fixed, since I haven't seen it since a couple of bug-fixes were added to that bit of code be someone, but I guess it seems not :-( It would be nice to be able to enable the intelligent AI by default, at least at level 1, for one of the next couple of releases.
Am I right in thinking there might also be an issue with crashes at the point of the "now clear of my airspace" message?
Melchior is refering to this
#0 0x080c9996 in FGAIPlane::GetLeg (this=0x0) at AIPlane.hxx:80
cptf #1 0x080c1c3f in FGTower::ProcessDownwindReport (this=0xd6bedb8, t=0xddccce8) at tower.cxx:664
cptf #2 0x080c5e95 in FGTower::Respond (this=0xd6bedb8) at tower.cxx:521
cptf #3 0x080c9734 in FGTower::Update (this=0xd6bedb8, dt= 0.033333333333333333) at tower.cxx:357
cptf #4 0x080a5372 in FGATCMgr::update (this=0xb371d90, dt=0.033333333333333333) at ATCmgr.cxx:167
cptf One can really rely on ATC crashing at the right moment. ;-)
(Somehow a plane occurs which contains no AI and is not ourselfs?)
I did the changes you are refering to, because there were a couple of STL problems hidden. I may have broken something so I did some heavy investigations into that and the most plausible cause seems to me that planes with identical callsigns have been around. I introduced some prints, but the problem didn't trigger any more for me. The patch removes some unused code, too.
ATC.patch
Description: Binary data