Bug: Psychic signalmen

[DrSuperGood]

Absolute Block signals logically work using telegraph equipment, or other communication systems. This is so that combined signals can know the state of the next signal box down the line.

However there comes a case where an absolute block network, with implicit telegraph/communication equipment, is interacting with another signalling network without such equipment such as plain time interval.

What should happen is that at the final stop signal of the absolute block network before the network without communication equipment the trains have to revert to drive-by-sight, after all the signal man cannot know if trains have passed a signal down the line or not. Instead what happens is that they somehow communicate with the trains psychically and so will only send a train in drive by sight if another train is in the block between them and the network without commication, otherwise they will reserve right up to the signal as if there were communications down the line.

The correct behaviour is to always send in drive-by-sight as they cannot know what the state of the signal box is ahead (no communications to...). Further more producing a warning message would be a good idea, as well as allowing one to upgrade railway policeman cottages, in order to inform the player that such an inefficiency is occurring. The really truly correct behaviour would be that when the train passes a signal with communication and no other trains have passed the signal then it can use block reservation behaviour to the signal between without communication, however handling this very specific case is not worth it especially since one will likely phase out all such non-communication signals anyway.

[Vladki]

I think that this should be a little-bit different.

For the simulation case we can assume that the last absolute-block signal (and its signalbox), has appropriate communication line to the next signal(box) of whatever type. If the next signal is e.g. time interval, we can assume that its signalbox is hybrid - it cannot communicate with the signalbox ahead, but can communicate with the signalbox in rear. Thus the signal is operated in time interval mode. There is no need for intermediate drive by sight area.

[jamespetts]

Vladki is correct - it makes more sense (and makes the implementation much more workable) for the signals to assume that the next signalbox is equipped with the necessary equipment to communicate with the previous signalbox.

[DrSuperGood]

Still there is a bug here somewhere... I have observed trains passing absolute block stop signals towards time interval signals in drive by sight mode. Sometimes they go in absolute block mode, other times in drive by sight.

Given the above intent, the only time this should happen is if the stop signal is before an end of signalling sign since that has no equipment to inform the signal operators that a train has passed the sign.

[jamespetts]

Can you upload a saved game in which this transitioning to drive by sight mode happens reliably at a specific time and place?

[DrSuperGood]

I will look into creating one over the next few days. Seeing how I was upgrading my network on the server, it is possible that the issue was entirely transient.

[jamespetts]

Can I check whether this can any longer be reproduced?