Path-based signalling preview and test version

[orudge]

Comments/suggestions welcomed...

File removed by request of Michael Blunck, since he has released an official signals set featuring PBS signals, as Patchman has noted.

[Aya]

Well, the DBSetXL only includes the German style semaphores, which is why it only has signals for traffic on the right hand side.

Kinda figured that was the reason. The British-style semaphores (I assume that's what the others are) would look rather odd in a German-style GRF package.

However, MB has now sent me an updated signals(w).grf, which I've added to the TTDPatch New Graphics page.

That's exactly what I was after. I can throw my version in the bin now. Kinda feels like I wasted my time doing that, but I guess I now know enough to do some new original graphics.

The TTDPatch distro is fairly well documented, but the graphics packages' documentation is scattered amongst the respective authors' websites. One thing I think needs to be made clear on the TTDPatch "New Graphics" page is which of the GRF downloads are the files you'll need to have installed to complement the features of TTDPatch which are enabled by default (i.e. if you don't have a ttdpatch.cfg file). These are generally GRF files with a GRFID of 0xFFFFFFFF (like gui, trkfound, nsignals, elrails). Perhaps canals should fit in this category too. I forget if it's enabled by default.

Or perhaps it might make more sense (with the authors' permission) to distribute these particular GRF files in the TTDPatch distro.

Back on topic, I noticed the behaviour for manually switching between PBS-signals and pre-signals is slightly different. In the ttdpatch.cfg file, there are 4 keys pertaining to signal behaviour:

presignals
extpresignals

pathbasedsignalling
miscmods.noautomaticpbsblocks

With both signal types, you get an option to enable the automatic creation of those signals with specified signal configurations ("presignals", and "pathbasedsignalling"), but with PBS-signals, if you want to be able to manually change them, you have to enable "miscmods.noautomaticpbsblocks".

Most of the time, I find the automatic signal-conversion to be correct, but sometimes I might like to manually set a signal to be path-based, which is not possible.

Would it be possible to change the switch behaviour of PBS-signals to be similar to that of pre-signals, by removing the "miscmods.noautomaticpbsblocks" key, and adding an "extpathbasedsignalling" key?

[Aya]

Did you ask MB before posting this? Those look rather like his signal graphics, and I wasn't under the impression he tended to approve of things like this.

File removed by request of Michael Blunck, since he has released an official signals set featuring PBS signals, as Patchman has noted.

Probably wise. I hope MB wasn't too offended by my posting that. I did credit him in the GRF file info as per the readme for DBSet. I appreciate the effort that has gone into TTDPatch, and felt it was fair for me to give something useful back to the community. That seemed like a GRF package that was commonly requested, so rather than harrass someone else to do it, I thought I'd do it myself.

MB doesn't visit these forums.

I guess now he does.

To any TTDPatch developers who might be listening: is it possible to write a GRF to redefine all 256 signals instead of just the 240 new ones?

[Patchman]

To any TTDPatch developers who might be listening: is it possible to write a GRF to redefine all 256 signals instead of just the 240 new ones?

Yes, replace the other 16 by an Action A.

[Patchman]

Alright, some pre-alpha-release replies...

Got a new crash in Alpha 45 :

The junction worked fine up to Alpha 44, and is constructed according the rules, so the crash shouldn't happen.

Unfortunately, that save was a bit too late, the train which the second one runs into has already unreserved that piece of track. I'd need a save from before that happens. Maybe it was actually due to an earlier bug in fact.

Ive caught a train in the act. Sorry havent got it before its happened. I think whats happened is a train has been waiting too long to leave the station, and reversed itself back into the "entrance" area then reserved itself into a platform with a train on it. You may need to pause it as soon as you load it.

That save was also too late, the train had entered the block already. To fix PBS bugs, I need savegames from before the train in question approaches the signal.

Here's one, where a train got bored, reversed, and crashed into another one.
Also contains cfg files.

I haven't been able to reproduce that, when I load the save there is no crash. Let me know if it still happens in a47.

I finally figured out why my trains were locking up!

It seems that at PBS blocks, the trains get bored too quickly. ... but in PBS blocks they don't wait even close to that long.

That's because you have "trains reverse at end of line and at stations" in your difficulty settings. That means, a train checks both directions when leaving a station, and if it can find a better path in the other direction, that's where it'll go. It has nothing to do with the signals in this case (except that the signal is red and the other isn't, so the other is better).

2 PBS Crashes, look on the right side of the Oil Rig. Same junction for both.

You have a signal immediately following a PBS signal, that's not allowed and will cause crashes. Actually, in a47 it'll simply never let a train through...

A weird crash in a45: it seems to be normal signals that I've set to PBS to trigger level crossings. A train breaks down and another one runs straight into the back of it, very near where the save is, in fact. And, unusually for me, I haven't broken one of the junction rules, AFAIK.

Sorry, couldn't reproduce it here either. Maybe I need to have the exact same version of the grf sets...

[krtaylor]

I finally figured out why my trains were locking up!

It seems that at PBS blocks, the trains get bored too quickly. ... but in PBS blocks they don't wait even close to that long.

That's because you have "trains reverse at end of line and at stations" in your difficulty settings. That means, a train checks both directions when leaving a station, and if it can find a better path in the other direction, that's where it'll go. It has nothing to do with the signals in this case (except that the signal is red and the other isn't, so the other is better).

Well, I've always used that setting, I much prefer it. Usually I have through-stations where some trains reverse, and other continue on. When the trains do a loop, if there's another inbound train, they'll lock up and stop. Before PBS, they wouldn't do that - they had more patience. Is there some way this can be adjusted? I do understand your explanation though, and it makes sense.

[Patchman]

Well, I've always used that setting, I much prefer it. Usually I have through-stations where some trains reverse, and other continue on. When the trains do a loop, if there's another inbound train, they'll lock up and stop. Before PBS, they wouldn't do that - they had more patience. Is there some way this can be adjusted? I do understand your explanation though, and it makes sense.

I'm convinced that it has nothing to do with PBS. It also has nothing to do with patience, if the train can't find a way out in one direction (because of a red two-way signal), it'll go the other way. It won't even try to wait at the red signal unless the other way is blocked too.

[krtaylor]

It won't even try to wait at the red signal unless the
other way is blocked too.

Yes, this is true. I did not previously observe that to be a major problem. Maybe it always was, but the trains did not give up and Stop, and so it wasn't nearly as noticeable. Is there any way to make it so the trains realize that they're just looping back, and don't do it; at least not until after they've waited the requisite time?

[Patchman]

It would be a major amount of work to do that, particularly since it's not in fact looping (returning to the same tile) but simply taking an alternative path.

[krtaylor]

Yeah... Well, it'd be nice anyway, but not too urgent I guess.

[SHADOW-XIII]

BUG? REPORT:

those 2 trains are loading coal ... it's station from their order however not their track (wrong side of station)

when I decide (or they will be full, however no they cannot be full because lack of coal) I force on of them to leave the station ... so he does ... but goes not the way it should ... screenie explains everything ... savegame included

[Patchman]

Do you have a savegame from before those trains enter the station?

[SHADOW-XIII]

i'm afraid not ... I will search for it ... I should post it today or i tuesday

[DaleStan]

Am I violating a junction rule, or is this a bug in PBS?

In 5 days, train 13 will need to reverse, but be unable to do so safely.

Trains entering the other side of Shorhill West do not have this problem.

[SHADOW-XIII]

ok ... this is game I found ... you can reproduce error manually
trains 127 and 103 are approaching the station but:
- put turning at the end of station so that later the will go back to the same station
- change order to other place before they stop at this station
- after train pass station but before pass junction and turns left change an order again to the same station (so that it will go right to those platforms)
- the best stop 1st train before second one approach to (and both platform as empty)
- when both trains stop at station .. you can change order of a train like I showed on screenie BEFORE

here is what to do

[Patchman]

Am I violating a junction rule, or is this a bug in PBS?

Was a bug.

ok ... this is game I found ... you can reproduce error manually

Ah, now I see, thanks. It was a bug with custom bridge heads. Bridges aren't supposed to give trains a choice where to go

[Mek]

Found another PBS bug... in alpha 46 i had a train crash, but in alpha 47 the second train involved in this crash will just wait forever (sometimes reversing, heading into the backside of a signal, re-reversing and stopping again) in front of a green signal, leading into a completely empty pbs block... The only thing that's special about this block is that it's a signal, followed by a tunnel, immediately followed by another signal...
(i know there are a couple incorrectly set-up pbs blocks in this savegame, that's because i played this game with a46, and they never caused any problems there. but this one should work as far as i can tell.. (surely because sometimes when train 45 momentarily leaves the tunnel-head/pbs-signal, some other trains get passed it...)

edit: wrong savegame attached

edit: played some more with alpha47, and it looks like it caused much more problems than were solved.. (at least in my case...) i just had like 5 trains crashing into each-other... if you open my second savegame, in a view days the first two trains will crash, and every time one of those trains gets cleaned up another crashes into the still existing remains...

[krtaylor]

An expansion on the PBS problem I mentioned before. When a train is in a terminal stub track (i.e. there's only one way in and out), if there's heavy traffic in the interchange block and it waits longer than it likes, it will give up and Stop. But there is no other way out, that's the only exit, it might as well keep waiting.

[Patchman]

edit: played some more with alpha47, and it looks like it caused much more problems than were solved.. (at least in my case...) i just had like 5 trains crashing into each-other... if you open my second savegame, in a view days the first two trains will crash, and every time one of those trains gets cleaned up another crashes into the still existing remains...

Well, I fixed one (as it turns out) relatively major bug, but I didn't notice some other code depending on the behaviour of the original, buggy implementation. So now I need to fix all that too. We're getting ever closer

An expansion on the PBS problem I mentioned before. When a train is in a terminal stub track (i.e. there's only one way in and out), if there's heavy traffic in the interchange block and it waits longer than it likes, it will give up and Stop. But there is no other way out, that's the only exit, it might as well keep waiting.

If a48 doesn't fix it, please submit a savegame and cfg that exhibits the problem.

[krtaylor]

After all these complaints about PBS problems, it's time for a compliment - the way the reserved path gets un-reserved as the train moves off of it, before it's even out of the block, is just wizard. I never thought it would get that good.