[railguy]

I've been working on creating an activity for BNSF Seligman (Open Rails), and have been making test runs. I'm running into an intermittent problem with the sim crashing. I've attached the log file. It does show various rolling stock issues, etc., but that does not appear to be the issue. At the bottom of the log, it describes an error with what I think is the precipitation module. I've also attaches a screen shot of the error message. The activity is crashing in heavy snow conditions. The activity has several weather change events set up as location events, but no changes are progress at the time of these crashes. They do seem to occur with some regularity about a minute or so after I do a save of the activity. I don't know whether or not this is a bug or some other issue.

I also have noticed the "shaking" that others have mentioned in V. 36. I've downloaded V 37, but have not installed it just yet. Any help appreciated.

Thanks.

Attached thumbnail(s)

• 

Attached File(s)

•  OpenRailsLog.txt (18.37K)
  Number of downloads: 270

[Csantucci]

This seems to be the same problem like this one
http://www.elvastowe...post__p__252163
so it does not seem to be caused by a specific route...

[railguy]

I'm wondering if it has to with elevation. This first line of the error message makes me wonder (my emphasis in italics):

at Orts.Viewer3D.PrecipitationPrimitive.HeightCache.GetHeight(WorldLocation location, TileManager tiles, SceneryDrawer scenery) in F:\ADRIANA\Carlo\OR_Work\Git_ORTS_source_mio\Source\RunActivity\Viewer3D\Precipitation.cs:line 446

The Seligman Route gets up to around 7,500 ft. elevation, and these crashes started occurring at somewhere around 6,000 ft. elevation and seem to get more frequent as the train climbs above that. I'm wondering if something in the code is causing an exception when the "elevation" of the tile mathematically exceeds what the precipitation module can compute for the "height" of the precipitation particles.

[Csantucci]

Thanks for the hint, but looking at the code it rather seems something related to indexing within the tile Height matrix. Does your route have tiles with dimension greater than 2048x2048m?

[railguy]

I don't know. It's an ummodified Streamlines BNSF Seligman (Open Rails) route.

[Csantucci]

I did two long runs in autopilot mode (Philly to Washington, Denver to Craig), but I didn't get crashes. I could insert checks in the code to avoid crashes, but I'd prefer to understand why the indices go out of bounds.

[railguy]

One additional note--in my activity, I have precipitationIntensity set at 0.020. My system is 64-bit and I've used this setting without problem before. Could that setting cause any issue with V36?

[Csantucci]

Rel 38 of OR NewYear MG logs some data just before the precipitation crash.
Railguy, if you run such release and get the crash, please attach the logfile as you already did: I should get useful data to understand how to avoid the crash.

[edwardk]

Did this problem happen with the earlier releases such as before 64bit? Just like the shaking wagon issue, it looks as if we are beginning to see issues that are connected with 64bit. My take on this is we could be seeing variable precision issues or something else that is involving the code and the 64bit process.

[railguy]

Csantucci, on 05 November 2019 - 10:13 AM, said:

Rel 38 of OR NewYear MG logs some data just before the precipitation crash.
Railguy, if you run such release and get the crash, please attach the logfile as you already did: I should get useful data to understand how to avoid the crash.

It will likely be a few days before I can get to it--I'm going to be away from my computer for most of that time. As soon as I do rerun the activity, I'll report on what I find.