Testing v8.004

[kapitaen]

The route also runs smoothly with OR version = 1.5.1.
However, this warning does not appear with fork v0.8.004a.

Screen Shot 04-16-24 at 08.06 PM.jpg

It only appears in the log as

[I] tsection out of sync !!! 

@ weichenpfleger, bei mir läuft die Strecke auch nach dem bearbeiten.

Jan

[eric]

The TDB synch warning showed up for me when I opened up Europe2 route using 8.004, and it should have shown up in lines 52-55 of your logfile:

capture_500618.jpg

If you have older copies of the TSRE logfiles, they'd show an update to TDB and tsection numbering. I don't see that here, which makes me ask if you'd already re-synch'd the database to match the TSection.

[weichenpfleger]

I had already synchronized the TDB, then opened it with 8.004 and saved it again. Then started OR and immediately got a crash. Then opened the route with 7.012, saved again and started OR - no crash.

[eric]

Without having logs to compare between your 8.004 save and your 7.012 save, I'm not sure how to pinpoint what's going on.

In 8.003 and even 8.004, TDB processing hasn't been touched, aside from disabling a NULL Entry compaction function that was being done at the time of a save, which happened in 8.003N.

I'd also question if this was due to an issue in build OR_T1.5.1-925-g1c557094e.

[eric]

OK, I was able to recreate weichenpfleger 's error

There's an issue with the track database saves that I'll have to work out. It's related to something that started happening in 8.003n so I'll have to spend some time diving into the root cause.

[eric]

v8.004b is available

This should address the TDB corruption issues which were reported with 8.004a and may be occurring within v8.003
Also adds "reload settings" function which will reload some (not all) of the values from Settings.txt


weichenpfleger kapitaen if you can, please confirm that you are able to add and delete track sections, and also open the route up in OpenRails after a save.


Suspected Root Cause: Gaps in the TDB TrackNode values (left after disabling "DeleteNulls" in v8.003) were occasionally resulting in the TrackNodes value stored at the top of the TDB being larger than the number of actual TrackNodes, and that was causing the Index Out Of Bounds error in OpenRails.

This only happened if there were tracks deleted from the TDB. Adding tracks alone did not appear to cause this.

[kapitaen]

Testing v8.004b,
Tracks in Europe2 deleted, driven in version 1.5.1 -
Added tracks in Europe2, driven in version 1.5.1 -
No Crash, no error.
The number and TrackNode contained matched in each case.

Jan

[weichenpfleger]

I can only confirm Jan's result! No problems with the start of OR!
Thank you, Eric, for your efforts to further develop TSRE5.

[eric]

Thank you both for identifying a critical flaw.

Sent from my SM-S911U using Tapatalk

[eric]

I lost a couple posts in error during some cleanup of the bug report thread.... I'm consolidating that here because it was spread across multiple threads.

• Derek noted the total mass error (metric tons to scientific notation)
• Martin was mentioning issues with changing track elevation

I'm looking into both.

docmartin if you could... try opening up your route in v7.012 and doing a save. Then come in using v8.004b and let me know if it persists.

[docmartin]

Did as requested and glad to say track issue disappeared in v8.003b. However, marker problem remains in v8.003b, at least for me. The markers are listed in the Navi Window and I can jump to them in TSRE but they do not all appear in TSRE.

Martin

[eric]

OK, that's a possible known issue.

For markers set by the editor (e.g. speedposts, sidings, platforms) the flags are showing up where expected, and the jump locations match.

For items in a MKR file.... the jump location is correct, but for some reason the marker itself is being rendered at its inverse Z value.

If the marker is at:

• position 100, 25, -100, the flag is is appearing at 100, 25, 100
• position 100, 25, 100, the flag is is appearing at 100, 25, -100

I'm fishing around the code to find where the missing pos/neg indicator is happening. It should be pretty easy to fix once I find it.

[eric]

Derek noted the total mass error (metric tons to scientific notation)

derekmorton I tested about two dozen consists tonight, and didn't find a single one with anything other than tons.

Can you perhaps post a consist I can then recreate and test?

docmartin I believe I've cracked the nut for the markers being out of place.... can you verify if they're showing up in your route at the reverse Z position like I mentioned above?

[derekmorton]

Hi Eric,

I have found the cause.
tsre gets confused when there are include files in the eng file.
I suggest it is already off the normal execution path.
The first sign is large tonnage.
If I keep going looking at other con files, it will eventually crash.
Log file is no help, last entry is the consist shown in the pic that I posted.

[docmartin]

can you verify if they're showing up in your route at the reverse Z

Yes, that indeed seems to be the problem.