The International Simutrans Forum

 

Author Topic: R3919 - pre-signal issue (possibly)  (Read 4231 times)

0 Members and 1 Guest are viewing this topic.

Offline jamespetts

  • Simutrans-Extended project coordinator
  • Administrator
  • *
  • Posts: 20700
  • Cake baker
    • Bridgewater-Brunel
  • Languages: EN
R3919 - pre-signal issue (possibly)
« on: October 29, 2010, 01:00:42 PM »
I notice that with the Standard nightly R3919 (same behaviour with Experimental), pre-signals do not behave in the way in which they did before (probably following the change to the block management system recently), with the result that, on the demo.sve for pak128.Britain (both the latest version and the second latest version), trains get permanently stuck at the pre-signal at the Christminster North station, whereas they did not before. I am currently unsure whether this is a bug with the current code, or a badly designed signalling system that only worked owing to a bug with the previous code that has now been fixed.

Offline Lmallet

  • Devotees (Inactive)
  • *
  • Posts: 727
  • 128 on the brain...
  • Languages: FR, EN
Re: R3919 - pre-signal issue (possibly)
« Reply #1 on: October 29, 2010, 01:27:49 PM »
I've had issues with pre-signals not evaluating the next choosesignal, causing jams.  As jamespetts said, I don't know if this is a bug or a feature.  :)

Offline Spike

  • *
  • Posts: 1361
  • First Simutrans Developer and Graphics Artist
Re: R3919 - pre-signal issue (possibly)
« Reply #2 on: October 29, 2010, 01:30:16 PM »
Prissi wrote in the German forum, that he is currently rewriting the presignal code in order to fix some bugs. Maybe this one was one of those. or maybe it was the result of the rewrite, which is not quite complete yet, as I understood.

Offline prissi

  • Developer
  • Administrator
  • *
  • Posts: 10530
  • Languages: De,EN,JP
Re: R3919 - pre-signal issue (possibly)
« Reply #3 on: October 29, 2010, 07:02:46 PM »
The presignals did simple not work as presignals, they only worked, when you cascaded two presignals and also did not work with a pre-signal followed by either a choose or long_block_signals. Now these (albeit seldom used combinations) should work. And there was still a bug with reversing convois, which reserved a tile too many under certain circumstances.

Offline Dwachs

  • DevTeam, Coder/patcher
  • Administrator
  • *
  • Posts: 4851
  • Languages: EN, DE, AT
Re: R3919 - pre-signal issue (possibly)
« Reply #4 on: October 30, 2010, 09:02:26 AM »
Does not work in 3920 too. Here is a simple setup causign the problem. Presignals are placed at station tiles.

Offline Dwachs

  • DevTeam, Coder/patcher
  • Administrator
  • *
  • Posts: 4851
  • Languages: EN, DE, AT
Re: R3919 - pre-signal issue (possibly)
« Reply #5 on: October 30, 2010, 01:05:43 PM »
Patch attached. Imho, the problem was that there was a presignal at the destination tile, which could not look ahead simply because the route ends at the signal.

Offline Erik

  • *
  • Posts: 219
  • Languages: NL, EN
Re: R3919 - pre-signal issue (possibly)
« Reply #6 on: December 04, 2010, 07:44:21 AM »
The presignals did simple not work as presignals, they only worked, when you cascaded two presignals and also did not work with a pre-signal followed by either a choose or long_block_signals. Now these (albeit seldom used combinations) should work. And there was still a bug with reversing convois, which reserved a tile too many under certain circumstances.

How is it with the progress of the reprogramming?

I'm building a .pak set of dutch train signals.
http://forum.simutrans.com/index.php?topic=6239.0
But i don't really understand how a pre_signal has meant to work in Simutrans.

Offline Dwachs

  • DevTeam, Coder/patcher
  • Administrator
  • *
  • Posts: 4851
  • Languages: EN, DE, AT
Re: R3919 - pre-signal issue (possibly)
« Reply #7 on: December 04, 2010, 08:12:53 AM »
The re-implementing is done.

Did you read the wiki about presignals?

Please ask your question in the help-request board again. Maybe searching the forum can answer this question, too.