Track Node not found in .tdb. TSRE writing incorrect entry in .tdb

[docmartin]

I have discovered that the reason OR crashed was because, after doing the corrections to the .tdb entry for the turnout section, the Track Node count at the head of the .tdb file no longer matched the number of Track Nodes listed in the file. Correcting this number appropriately allowed OR to load properly. Even so, however, the warning message about not finding Track Node 2832 in the .tdb file still appeared in the log file.

[kapitaen]

I had already suspected that it was not the turnout shown.

Place the camera vertically, go exactlyl to this position

-6080 14925 x 891.311 z 558.573

and take a picture. I would like to see the position in the navigation window, please.

Jan

[docmartin]

Jan,

Here is the picture you want. I had to reduce the size to meet the limit on uploads but presumably you can save the image and then expand it so you can see the detail. It proves that the turnout in question is in fact the turnout with UiD 2832.

TDBproblem5.jpg
By the way, do you know if it is possible to specify locations in the navi wiindow with three places of decimals? I don't see anything in the settings file allowing you to do this.

Martin

[eric]

If it's not affecting routing trains thru the turnout, I'd just ignore it.

If it's affecting train routing...

1. Delete the turnout and any tracks connected to it
2. Save, exit and re-open
3. If there are still yellow lines on the turnout, there is a ghost copy of the tracknode somewhere in the TDB
4. Attach a 5m section of track to the red post, and use the delete TDB vector hack twice to wipe out both vectors
5. Save, exit and re-open
6. Re-lay the pieces you deleted

Because TSRE holds all the pieces in memory **including deleted pieces** it's important to save, exit, and re-open when you're running into TDB oddities.

[eric]

By the way, do you know if it is possible to specify locations in the navi wiindow with three places of decimals?

Can I ask why?.... The Navi Window controls the camera. I can't see a case for needing camera precision less than a meter... It's the pointer that gives you centimeter precision if needed.

[kapitaen]

Thanks for your picture.
Eric has actually said everything you need to know.
Unfortunately I can't recognize the ID in the picture even with 10x magnification. If this is 38668, then it is a ghost track. If not, then there are two tracks on top of each other.

Jan

[docmartin]

First, I would like to thank both of you, Eric and Jan, for having spent so much time on my problem

I tried following the instructions for removing the ghost tdb entries. I deleted all the relevant track sections and was left with the turnout yellow lines, one red pole and two blue poles as you would expect. Attaching a short straight section to the red pole, it would only face away from the turnout and hacking simply removed the tdb vector for that section of track, not for the ghost turnout. I could get rid of the blue poles, however, using the same procedure i.e. attaching a straight section to each blue pole in turn and then hacking. However nothing would shift the red pole. I even tried directly editing out the red pole from the track database. This worked to the extent that on opening TSRE all the unwanted poles has disappeared and I was able to replace the track sections I had removed originally. However, OR crashed on loading with the usual Fatal Error message. So it looks as though there is no way of eliminating the ghost tdb entries.

Now, as far as I know if I leave well enough alone, there is no effect in OR. It runs OK and it is even possible to create paths over the suspect track section. So, as Eric says, I will probably have to ignore the problem and simply live with the fact that I am getting warning messages in the log file. Not very satisfactory but at the moment I see no alternative.

As for the three decimal places in the navi window locations, I was just curious since Jan in his request had specified the location as given in the tdb file, i.e. to that number of decimal places. I absolutely agree that there is no need for that level of precision in the navi window.

Martin

[kapitaen]

I wouldn't give up just yet. Can you upload the route somewhere?

Jan

[docmartin]

Jan, You can download the route from GitHub. Or if you have the latest Unstable version of OR, then you will see that there is a new "Content" button on the opening screen. This will open a menu of routes you can download and install immediately, including my Chiltern route.

Martin

[eric]

I'll try to pull it down as well.

This may be an argument for not distributing routes via GitHub... Having just one file out of synch can be anything from annoying to catastrophic.

Is there a big difference between what is on GitHub and what was already in the ukts library? I've bundled the route along with all of the cumulative patches here:

https://www.trainsim.com/forums/filelib/search-bundle?bundle=39520

[kapitaen]

All right, here it comes. We have the following error.

Warning: {TileX:-6080 TileZ:14925} track node 2832 could not be found in TDB

in WorldFile:
TrackObj (
UiD ( 2832 )
SectionIdx ( 38668 )
Elevation ( 0 )
JNodePosn ( -6080 14925 891.311 28.5577 558.573 )
CollideFlags ( 39 )
FileName ( ukfs_tl_17m5_9d46.s )
StaticFlags ( 00200180 )
Position ( 891.311 28.5577 558.573 )
QDirection ( 0 0.56912 0 0.822254 )
VDbId ( 4294967295 )
)

in TSRE:
Screen Shot 04-02-24 at 08.53 PM.jpg

But the search pattern 38668 -6080 14925 2832 cannot be found in the .tdb.
Then the error must be in the WorldFile, as the TSRE does not report any errors.
We now use a search pattern that works with the position, as there are many switches in the WorldFile.

We search with

-6080 14925 891.3

no longer specify decimal places.

For the matches we check the presence of the UID in the WorlFile. If it is present, we search further.

Screen Shot 04-02-24 at 09.03 PM.jpg

If there is no match in the WorldFile, then we have found the culprit.

Screen Shot 04-02-24 at 09.03 PM 001.jpg

With this value 7356 we now replace the UiD ( 2832 ) to UiD ( 7356 ) and save the file.

In the TSRE it now looks like this

Screen Shot 04-02-24 at 08.55 PM.jpg
and the error is fixed.

You can now repair the following errors yourself...

Jan

Warning: {TileX:-6082 TileZ:14925} track node 12986 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 16757 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 16758 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 16759 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 16760 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 26316 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 21060 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 21061 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 21062 could not be found in TDB
Warning: {TileX:-6082 TileZ:14925} track node 21063 could not be found in TDB​

PS: someone could fix your shapes, it's horrible to look at.

[docmartin]

Thanks, guys.

Is there a big difference between what is on GitHub and what was already in the ukts library?

Yes, there is a lot of difference. I was persuaded to use GitHub as, being a work in progress, it offers a pretty foolproof way for people to stay abreast of the latest iteration of the route. Since the last version available on UKTrainSim I have added a lot of track and a lot of scenery.

someone could fix your shapes, it's horrible to look at

What are referring to here, Jan?

OK, Now I am going to spend some time trying to replicate Jan's solution. Wish me luck!

Martin

[eric]

docmartin I've added a link in our file library to the new version on GitHub.

[docmartin]

Thank you for this, Eric.

Thanks to your and Jan's help I have now successfully eliminated all the Track Node errors. Interestingly, some were caused by the Track Vector in the .tdb file referencing a UiD which did not exist in the WORLD tile and some by the Track Vector referencing a UiD which did exist in the WORLD tile but which related to a Track Object rather than a Track Section, e.g a speedpost.

Martin