The International Simutrans Forum

Simutrans Extended => Simutrans-Extended development => Topic started by: scamps on November 05, 2013, 01:14:39 PM

Title: Long block signals behave differently in Experimental?
Post by: scamps on November 05, 2013, 01:14:39 PM
I tried the "put 3 stations between long signal loops" thing, which works in Standard. It did not work out in Experimental.

Attached two test saves - Standard (pak Britain) and Experimental (11.12).
https://www.dropbox.com/sh/7xsv2xemfdc6jwc/ANDUQOXc-t (https://www.dropbox.com/sh/7xsv2xemfdc6jwc/ANDUQOXc-t)

If long block signals sould behave differently in Experimental, where can I find correct description?

Thank you.
Title: Re: Long block signals behave differently in Experimental?
Post by: whoami on November 05, 2013, 04:30:36 PM
Perhaps the same bug as discussed here: http://forum.simutrans.com/index.php?topic=12393.msg126489#msg126489 (http://forum.simutrans.com/index.php?topic=12393.msg126489#msg126489)
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on November 13, 2013, 03:59:09 PM
bump?

Edit: @whoami
OK, I followed the link you provided and thought that the mentioned bug was fixed prior to 11.12. Maybe I have misunderstood.

Editing post not to bump the thread again.
Title: Re: Long block signals behave differently in Experimental?
Post by: whoami on November 14, 2013, 02:49:02 AM
Hehe, I am going to bump you back.  ;)
Uhm, I mean: We have to wait for the next ST-exp version.
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on November 17, 2013, 06:51:41 AM
New behavior in 11.13: now even one train can not travel this line, "wating for clearance"
Title: Re: Long block signals behave differently in Experimental?
Post by: whoami on November 17, 2013, 04:39:16 PM
For this, there might be a simple explanation: sometimes (also in ST-standard), especially on loading a savegame, an invalid reservation remains on track tiles. Use 'b' to view and fix the reservations.
Title: Re: Long block signals behave differently in Experimental?
Post by: Carl on November 17, 2013, 04:59:35 PM
Long blocks look to be working fine for me in 11.13.
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on November 17, 2013, 05:34:56 PM
Then it means I use signas in some wrong way. After upgrading to 11.13 I made a new line, but it made no difference. Train stands at long block signal, does not reserve anything and waits for clearance.

I would appreciate if someone checked out the save and explained what is wrong with this cofiguration.
Title: Re: Long block signals behave differently in Experimental?
Post by: Carl on November 17, 2013, 05:40:44 PM
I'm not able to look at the save right now, but this reminds me of a problem I had a few months ago. Does this help? http://forum.simutrans.com/index.php?topic=11749.msg115576#msg115576
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on November 17, 2013, 07:04:30 PM
Useful info, Carl, but does not seem to be the case.
Title: Re: Long block signals behave differently in Experimental?
Post by: jamespetts on November 18, 2013, 09:43:36 PM
Carl - thank you for testing.

Scamps - if you are still having problems, can you upload a saved game in which you can reproduce something that is clearly incorrect, and describe the incorrect behaviour precisely? That would be very helpful.
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on November 19, 2013, 02:52:45 AM
First post contains a link to two saves. One Standard game, one Experimental game. Same track confguration.  In Standard two trains can use the line, in Experimental 11.13 a single train can not use the line, "waiting for clearance" at long block signal.

And repeating the link   https://www.dropbox.com/sh/7xsv2xemfdc6jwc/ANDUQOXc-t

edit: failed to include screenshot directly in post, added under the same link.
Title: Re: Long block signals behave differently in Experimental?
Post by: whoami on November 19, 2013, 12:58:19 PM
I tested your savegames (I did not before). The funny thing is: the savegame from ST-std works fine in ST-Exp 11.13! signal-exp started to work on the left end (only) once I added signals on the bypass tracks and removed the "mirror" flag from the line schedule and the "reverse" flag from the trains. signal-std has a circular schedule, but changing it to a bi-directional schedule (still in ST-Exp) does not prevent it from working.
In signal-exp, I also extended the stations in a way that train length should not matter, but it did not help either. Forcing the trains to proceed or return to the previous stop does not prevent the problem from reappearing. No stale block reservations are visible.
The differences between the two savegames are schedules, train and station lengths, vehicle types, also slight differences in track layout. I still cannot exclude any of these as the reason for the misbehaviour.
Title: Re: Long block signals behave differently in Experimental?
Post by: jamespetts on December 22, 2013, 12:10:57 AM
Hmm - perhaps I have misunderstood the intention of long-block signals when applying changes to the routing and block reservation systems. Is the idea to allow one train to follow another into the same section, separated only by station stops rather than explicit signals?
Title: Re: Long block signals behave differently in Experimental?
Post by: TurfIt on December 22, 2013, 01:11:30 AM
No, rather the opposite. Long block should see through stations so reversing trains don't deadlock.  Let's you implement a 'One Train Working' system on a single track branch with multiple stations.
Title: Re: Long block signals behave differently in Experimental?
Post by: jamespetts on December 22, 2013, 10:58:21 AM
Hmm - so how should this example behave, in that case?
Title: Re: Long block signals behave differently in Experimental?
Post by: scamps on December 29, 2013, 03:47:41 AM
I used this as an example on long block signals
http://forum.simutrans.com/index.php?topic=10472.0 (http://forum.simutrans.com/index.php?topic=10472.0)
Title: Re: Long block signals behave differently in Experimental?
Post by: jamespetts on January 11, 2014, 09:41:13 PM
I have just fixed a bug relating to long block signals that would in some cases cause convoys to become stuck behind a long block signal unnecessarily, which I hope will fix this issue.