Elvas Tower: Timetable Argument out of range Exception - Elvas Tower

Jump to content

Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic

Timetable Argument out of range Exception Timetable works in x4001 but unreliable since about x4075 Rate Topic: -----

#1 User is offline   rickloader 

  • Conductor
  • Group: Status: First Class
  • Posts: 479
  • Joined: 05-February 13
  • Gender:Male
  • Location:Southampton uk
  • Simulator:Open Rails
  • Country:

Posted 18 March 2019 - 02:23 AM

Since about the New Year my timetable has failed with an Argument out of Range exception.I`m posting this in the bug section because the timetable loads in x4001, but is increasingly unreliable after about x4075 and will only run early trains in the x1.3.1-31 versions.
Attached open rails logs.
I am sorry about the number of errors. Obviously they don`t help, but please don`t berate me about errors : instead some suggestions how to reduce them would help. The fact remains the timetable works better in earlier OR versions.
Over the last 3 months I have rebuilt the timetable,checked paths, platforms, signals. reduced sound errors. checked the route with route riter and tsutils "clrdb", but not enough expertise to tackle the tdb direct.
A particular problem is lack of consistency. By this I mean a train tests ok, so moving on to the next train I find the first no longer runs correctly. So I tweak paths, syntax, signals, then back to initial settings and it works : only to find another train misbehaving !
$transfer and $detach seem most affected. Also I don`t understand "multiple pickup at location" errors.
After testing and discarding hundreds of timetables, I would appreciate any guidance.
Thanks, rick

Attached File(s)



#2 User is offline   roeter 

  • Vice President
  • Group: Status: Elite Member
  • Posts: 2,420
  • Joined: 25-October 11
  • Gender:Male
  • Country:

Posted 18 March 2019 - 02:24 PM

View Postrickloader, on 18 March 2019 - 02:23 AM, said:

Since about the New Year my timetable has failed with an Argument out of Range exception.I`m posting this in the bug section because the timetable loads in x4001, but is increasingly unreliable after about x4075 and will only run early trains in the x1.3.1-31 versions.


That's very strange as there have been no changes to the timetable mode since June 2017 (version 3875). The main module, TTrain.cs, has seen only one change committed since that version, and that change was only about a call to a method which needed changing as the 'parent' method had changed. It looks like something else is causing these problems, but I cannot tell what this could be.
I have problems placing the crash - I'm not sure I'm looking at the correct line.

Regards,
Rob Roeterdink

#3 User is offline   rickloader 

  • Conductor
  • Group: Status: First Class
  • Posts: 479
  • Joined: 05-February 13
  • Gender:Male
  • Location:Southampton uk
  • Simulator:Open Rails
  • Country:

Posted 19 March 2019 - 02:33 AM

Thanks very much for your reply Rob. A couple of questions please.
Would the "tt output full evaluation" give any useful information?
Could "multiple pickup for same location" relate to the TSRE5 platform length bug you highlighted on trainsim.com?. If so I could delete and replace platforms as Goku claims to have fixed this in the latest TSRE.
Might changes in Path handling explain the results in different OR versions? If so I could try different OR versions to pinpoint where the tt problem starts
A strange part of the problem is that the timetable plays differently when observed, to when the trains are run internally.
Examples.By this I mean that when a certain detaching train is player train or seen from an observer train, the $detach works ok. However, when the detach is run unseen by computer it does not occur.
Conversley I spent ages trying to get a $transfer to work and the active train would not see the passive train as a train to couple to. Imagine my surprise when the passive train later passed complete with the transferred carriages. The transfer does complete off screen.
Perhaps these are just examples of my timetable inconsistency. I did wonder if the timetable collumns not being arranged in strict time sequence might confuse OR. However a timetable reconstructed with collums in time sequence still failed.
Just in case I have done something silly in the time table, here it is attached. Pls view in open office or similar.
I can`t easily share my wip route as it is rather big.

Attached File(s)



#4 User is offline   rickloader 

  • Conductor
  • Group: Status: First Class
  • Posts: 479
  • Joined: 05-February 13
  • Gender:Male
  • Location:Southampton uk
  • Simulator:Open Rails
  • Country:

Posted 20 March 2019 - 01:24 PM

I`m going to overhaul the route as a possible cause of timetable inconsistency. I wish I had more expertise. I may be gone some time......................!
First step run tsutils. Unfortunately, ts utils adds a comment which OR does not seem to like. I will have to edit every world file. This tile also seems odd as it only has a view sphere.


SIMISA@@@@@@@@@@JINX0w0t______

Tr_Worldfile (
VDbIdCount ( 1 )
ViewDbSphere (
VDbId ( 0 )
Position ( -79.9147 18.4118 -754.418 )
Radius ( 784.919 )
)
)
Comment ( "TS DB-Utility - Version: 3.6.05(01.04.2012), Filetype='World', Copyright © 2003-2012 by C.H.Rave" )

#5 User is offline   roeter 

  • Vice President
  • Group: Status: Elite Member
  • Posts: 2,420
  • Joined: 25-October 11
  • Gender:Male
  • Country:

Posted 21 March 2019 - 02:15 AM

Sorry for the delay in responding to your questions but I was out of reach of the internet for most of last week, and had a fair few things to catch up when I got home again.
Anyway, here's my thoughts on this.

View Postrickloader, on 19 March 2019 - 02:33 AM, said:

Would the "tt output full evaluation" give any useful information?

Not really, that's more intended for when there are conflicts between trains, causing delays and lock-ups.

Quote

Could "multiple pickup for same location" relate to the TSRE5 platform length bug you highlighted on trainsim.com?. If so I could delete and replace platforms as Goku claims to have fixed this in the latest TSRE.

No, that's not the cause. At worst, the incorrect platform definition is TSRE could cause the trains to stop at an incorrect position, but it would still be treated as that station stop.
As for the problem with these "multiple pickup" reports, I'll get back to that later.

Quote

Might changes in Path handling explain the results in different OR versions? If so I could try different OR versions to pinpoint where the tt problem starts.

As far as I know there are no changes to path handling which affect timetables.

Quote

A strange part of the problem is that the timetable plays differently when observed, to when the trains are run internally.
Examples.By this I mean that when a certain detaching train is player train or seen from an observer train, the $detach works ok. However, when the detach is run unseen by computer it does not occur.
Conversley I spent ages trying to get a $transfer to work and the active train would not see the passive train as a train to couple to. Imagine my surprise when the passive train later passed complete with the transferred carriages. The transfer does complete off screen.
Perhaps these are just examples of my timetable inconsistency. I did wonder if the timetable collumns not being arranged in strict time sequence might confuse OR. However a timetable reconstructed with collums in time sequence still failed.

There are some known problems when the player train is involved in attach or transfer actions. Corrections for these problems are 'in the pipeline' (it just takes a bit longer to get these through to the end of that pipe than I had anticipated).
Otherwise, it's rather strange that there would be differences in what happens when the trains are observed or not, as obviously it's all the same code.

Quote

Just in case I have done something silly in the time table, here it is attached. Pls view in open office or similar.
I can`t easily share my wip route as it is rather big.


I looked through the timetable but I did not see anything out of the ordinary.
That brings me to the problem of the "multiple pickup" errors.
It looks like you stumbled on a program error here - the pickups are not at the same location and should have been handled normally.
I will have to look into that, but that will take time. This could cause some of the problems you experience as the train might get stuck where it is supposed to the one of the pickups.
I would suggest that for now, you split the affected workings into two separate trains. Best thing to do is to terminate the train at the location of the first pickup, and define the pickup for that location in the #dispose field. The second pickup can remain as it is.
You can use the present full path for the second train but you will need to define a new path for the first portion as that path has to terminate at the first pickup location.
Could you please try that and let me know if it helped?

Regards,
Rob Roeterdink

PS. I seem to have picked up a rogue attachment. Whatever it is, it is not mine so please ignore it.

Attached File(s)



#6 User is offline   rickloader 

  • Conductor
  • Group: Status: First Class
  • Posts: 479
  • Joined: 05-February 13
  • Gender:Male
  • Location:Southampton uk
  • Simulator:Open Rails
  • Country:

Posted 15 April 2019 - 03:20 AM

Thanks for your help Rob. I tried splitting the "multiple pickup for same location", but then timetable would not load at all. It turns out these errors are a symptom not the cause ,as now the timetable is corrected they have disappeared. I dont think they are a "progam error"

So what was the cause? A faulty path. Now deleted and replaced. So why has it taken me 4 months to find? Read on and despair at how muddled rick can be!
1) The culprit 0036 Northan Goods spends 5 hours shunting en route as Brockenhurst yard pilot. It resumes at 06:15 split, and $formed as a second train. As the timetable passed 06:15 and only crashed around 08:10 I did not suspect this train.
2) When tested as player train both the 00:36 and the 06:15 ran ok.
3) I completely re-wrote the time table, and again the Northam Goods passed tests . New time table failed around 08:00 as before.
4)In the OR log there is a very useful hourly time stamp against time table errors. This led me to suspect the problem was after 06:00. It would be very helpful if the exact times were logged with errors.
5)I commented out all trains after 06:00 by placing #comment at the head of each column. Then I reintroduced trains to the timetable, testing each addition. The timetable failed with the 06:15
6) stages 1-5 could have been eliminated. I put the paths into date order and the 06:15 path was last modified in Dec18 just when the timetable became unstable.
So what makes a faulty path? Both track viewer and routeriter say it is ok. Maybe it relates to a bad .tdb such as "incomplete crossover", but I don`t know how to locate these.
attached is the faulty path.
Well, I`m delighted the problem is resolved. Timetable mode allows the reconstruction of operations from original 1950s documents, and I hope to complete a full day. It is very satisfying to see the whole network running. Thanks Rob!
rick

Attached File(s)



Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic

1 User(s) are reading this topic
0 members, 1 guests, 0 anonymous users