[roeter]

Looks to me this method can only loop if there is a real track 'circle' - i.e. the track-layout is such that if you follow the route you sooner or later end up at the same signal in the same direction - without ever finding the signal you are looking for.

Where about on the route does this happen?
I know there are some weird and complex layouts in this area (York, Lincoln, Doncaster), but I do not think any of those have an actual circular loop.

Regards,
Rob Roeterdink

[dennisat]

roeter, on 05 March 2014 - 01:05 PM, said:

Looks to me this method can only loop if there is a real track 'circle' - i.e. the track-layout is such that if you follow the route you sooner or later end up at the same signal in the same direction - without ever finding the signal you are looking for.
Where about on the route does this happen?
Rob Roeterdink

It happens at Little Steeping, heading towards Skegness. There are no circles in that direction but in the reverse direction there are a couple of return loops which may help to make a circular path but they are some distance behind the train.

Dennis

[dennisat]

dennisat, on 06 March 2014 - 01:05 AM, said:

It happens at Little Steeping, heading towards Skegness. There are no circles in that direction but in the reverse direction there are a couple of return loops which may help to make a circular path but they are some distance behind the train.

In case it's any use, the OR Log is attached. The program line numbers will not correspond exactly to yours because of all the Trace statements I had to insert. Also, I soon found out that adding the Trace statements initially paralysed the Update process because of the number of iterations these signal methods go through. I therefore had to add a little logic to track only the signal where the problem occurred.

Dennis
 OpenRailsLog.txt (263.12K)
Number of downloads: 217

[dennisat]

dennisat, on 05 March 2014 - 12:04 PM, said:

...I'll do a bit more research tomorrow. At the moment I'm starting from a save to do the tests. I'll start the act from scratch and see if the loop happens again...

Well I started it from scratch and the loop did not occur. There's something time critical about this - trains in slightly different places so that signals have different states, perhaps.
I've left the trap in in case the loop occurs in any other activity I might run.

Dennis

[roeter]

Just wonder - did you make the save when traveller.cs wasn't patched yet for the curve at Firsby?
Going forward from Little Steeping would lead to this curve, the problems with that track section might have corrupted the signal data as the signalling also uses the traveller to determine the position of the signals.

Regards,
Rob Roeterdink

[dennisat]

roeter, on 06 March 2014 - 12:14 PM, said:

Just wonder - did you make the save when traveller.cs wasn't patched yet for the curve at Firsby?
Rob Roeterdink

Yes. I had turned the Track Monitor off after it had started crashing earlier in the activity. I had continued until the hang occurred at Little Steeping when I restarted from a save at Boston and moved up to Little Steeping and saved a couple of hundred metres from the signal that triggered the loop. I then investigated the Track Monitor crash before the hang but always started the tests from the Little Steeping save. In retrospect I should have started the activity from scratch after locating the problem with the Track Monitor and checked whether the hang still occurred. I would have saved quite a few hours of work chasing that down. A Learning Opportunity is what they call that, I think.

Thanks for your interest.

Dennis

[roeter]

dennisat, on 06 March 2014 - 12:56 PM, said:

Thanks for your interest.

Dennis

Thanks for your help in sorting this one out. Would have been hard without actually having the route.

Regards,
Rob Roeterdink