(My) Problem on Long Block Signal

[Sholong]

Constructed a underground railway like what shown on the above image, I figured out that I should use two long block signals on each of the either side of the double-railed section, as to run the line with two trains while saving money by not to build double-platformed station. But eventually the long block signal seems to be acting exactly like a ordinary rail signal and causes my train coming from opposite direction at the same time to be jammed.


Is that I misunderstood the meaning of long block signal? cause in my mind, it is for reserving the whole block of rail until the next signal. I searched on the net and found some people stated that the next stopping station will act as a signal and end reservation making by the long block signal; However, some tutorials do not state stopping stations as rail signal, and this confuse me. If, it is really the case that the stopping stations will prohibit the effect of long block signal beyond it, does this mean that I will have no choice but i must use double-platformed stations, or at least build dual tracks between any consecutive single-platform station? Or else, can this be a bug?


Don't know if I should raise the question here in Experimental branch or in the Simutrans general discussion branch, I am using Simutrans 111.0 Experimental 10.11 -r1011 with the pak.128Britain 0.8.3, which the rail system was freshly constructed with the above version.

[AP]

I also have trouble with Long Block signals. I used to be able to make them work with 2 successive single-platform stations within a block, but after coming back to Simutrans after a break, even that gives me difficulties. I recall being told that changes were made to certain signals at some point in the past, always assumed it was due to that.

One such station in a block works okay, two or three ends up as you've shown (I think 3 was always too many). I find that having a "home" signal just inside the entrance to the passing loops, at each end of the single track section sometimes helps, but not always. I've also tried using presignals in a few variations.

If someone could clarify the "correct" way to signal one, two, and three single-platform-stations between loops that would be awesome.

[jamespetts]

Thank you for your report, and sorry that you are having trouble. I do not think, however, that this is an Experimental specific issue, since the code for long block signals is the same in Experimental as it is in Standard. The issue is, I think, as you identify, that both the first and second station count as signals if the trains stop there. There is in Simutrans, in effect, no way of having single line tokens.

Edit: Moved to help requests.

[Dwachs]

The long block should work as you described. So there could be a bug hidden. For ease of debugging, can you upload please a small sample savegame that demonstrates this false behavior?

[prissi]

I could be again a problem during saving reloading, depending on the simutrans version used.

[jamespetts]

Hmm, if this is a bug, it should perhaps revert to the Experimental section, although it would be helpful to check whether the latest Standard nightly displays this behaviour or not.

Incidentally, I tried and failed earlier to-day to integrate the recent update to Standard that involves saving and loading the next stop index. I couldn't get it to work, as it would sometimes save a number too high, causing a crash on loading when it tried to read a route tile that did not exist in the vector. Is this update related to this problem at all? If it is, I should be interested in any tips on integrating into Experimental.

[kierongreen]

Long block signals have been working as intended for me in recent versions.

[jamespetts]

Kieron - how recent, may I ask?

[kierongreen]

As in last 2 months, builds from svn.

[jamespetts]

Ahh, that explains it - Experimental isn't synchronised to that point yet. This was a bug in previous versions of Standard, then, before that?

[kierongreen]

Maybe? Look through svn log to check.

[kierongreen]

r5575 | prissi | 2012-03-17 21:19:08 +0000 (Sat, 17 Mar 2012) | 1 line

FIX: reservation and next stop now saved with game


and, in the event you didn't pick up on it at time:

r4392 | prissi | 2011-04-14 22:17:53 +0100 (Thu, 14 Apr 2011) | 1 line

FIX: long block signals working again


may affect this

[prissi]

Something is screwed with reservations in experimetnal anyway. Railroad crossings in the yoshi testgame stopped working the cars never obeyed the preparation for closing. (just scroll down to bottom one near the start position).

[jamespetts]

Hmm - I haven't changed crossings specifically in Experimental. A brief test with crossings shows that cars will generally obey them, but occasionally get stuck in the middle when the barriers close - I assume that that is what you mean by the preparation for closing? Also, the crossing does not close immediately that the track is reserved; is that the intended behaviour in Standard? How exactly ought it work?

[Sholong]

replicated the problem with a small save, hope it can help


created with Simutrans 111.0 Experimental 10.11 -r1011 with the pak.128Britain 0.8.3