[cjakeman]

ostkamp, on 01 November 2012 - 02:09 PM, said:

I can retry with any tweaked OR version and look up file contents anytime, if that provides the necessary info for you.
You just need to drop me a link to the changed version or patch.

I found a way to generate a similar error on one of the 6 MSTS routes and then tweaked the simulator to deliver a more specific message.

I've sent you a private message with a link to this version. Please try it on your commercial route and send me the log file so we can see how well it works

Thanks for your help.

[ostkamp]

Hello,

cjakeman, on 02 November 2012 - 02:31 PM, said:

I've sent you a private message with a link to this version. Please try it on your commercial route and send me the log file so we can see how well it works

Thanks for your help.

I just did it. I could not find RETURN statements at the mentioned line numbers but there was some reccurring pattern, please find the details in private message.

Regards

Guido

[cjakeman]

ostkamp, on 02 November 2012 - 02:59 PM, said:

I just did it. I could not find RETURN statements at the mentioned line numbers but there was some recurring pattern, please find the details in private message.

Thanks, Guido.

Just wanted to announce that a solution providing accurate line numbers and no repetition is not straightforward but Rob Roeterdink has offered to look into it. Thanks, Rob.

[James Ross]

cjakeman, on 02 November 2012 - 12:45 PM, said:

Thanks, James.

The following would be compatible, then, and now shows OR's position regarding the message.

Information: Not implemented - skipped notch type EngineBrakesControllerNeutralhandleOffStart in C:\Programme\Microsoft Games\Train Simulator\trains\trainset\GR-BR 01_DB\01_1057.eng:line 433

Warning: Bad data - cannot parse the constant number -0.04-2.1035 in c:\programme\microsoft games\train simulator\routes\rollbahn2\shapes\au1000_xx.sd:line 6

To be consistent with the dozens of other messages they should be:

Warning: Skipped unknown notch type EngineBrakesControllerNeutralhandleOffStart in ...

Warning: Skipped invalid constant -0.04-2.1035 in ...

But, as I said, I think the unknown notch type one should really be information as there's no data error (although I guess it might be a typo too...).

[cjakeman]

James Ross, on 03 November 2012 - 06:43 AM, said:

To be consistent with the dozens of other messages they should be:

Warning: Skipped unknown notch type EngineBrakesControllerNeutralhandleOffStart in ...

Warning: Skipped invalid constant -0.04-2.1035 in ...

But, as I said, I think the unknown notch type one should really be information as there's no data error (although I guess it might be a typo too...).

Yes, I agree with you that all the messages should be consistent.

I was volunteering to change all the dozens of messages so they indicate OR's position regarding the message. I.e this warning message is identifying bad data and that information message is identifying some feature that OR may (or may not) implement in the future.

I'm not being critical of the work done in error handling and I've found these pinpoint messages to be very useful indeed. Bit if we can minimise misunderstandings, that helps OR and saves us all time in the long run.

[James Ross]

cjakeman, on 03 November 2012 - 08:17 AM, said:

I was volunteering to change all the dozens of messages so they indicate OR's position regarding the message. I.e this warning message is identifying bad data and that information message is identifying some feature that OR may (or may not) implement in the future.

Okay, well, I don't like the proposed message changes that much but I'll support changing anything where it is skipping unknown data to be Information instead of Warning (and worded the same).

One reason is that "Skipped unknown <type> <name>" is what actually happened where as saying it is not implemented is making an assertion that we know about it. Ideally, all known but unimplemented items would be "Information: skipped not implemented <type> <name> in <location>" and typos would be "Warning: skipped unknown <type> <name> in <location>" but I don't know that we can really do that well everywhere. Thus, I'd prefer to keep the message stating only what we know - that it was skipped because it is unknown (to Open Rails at this time).

[cjakeman]

James Ross, on 03 November 2012 - 08:35 AM, said:

One reason is that "Skipped unknown <type> <name>" is what actually happened where as saying it is not implemented is making an assertion that we know about it. Ideally, all known but unimplemented items would be "Information: skipped not implemented <type> <name> in <location>" and typos would be "Warning: skipped unknown <type> <name> in <location>" but I don't know that we can really do that well everywhere. Thus, I'd prefer to keep the message stating only what we know - that it was skipped because it is unknown

Makes sense. Thanks for spelling that out so clearly.

I'll do my best with it.

[cjakeman]

cjakeman, on 03 November 2012 - 03:18 AM, said:

Thanks, Guido.

Just wanted to announce that a solution providing accurate line numbers and no repetition is not straightforward but Rob Roeterdink has offered to look into it. Thanks, Rob.

Rob has committed a solution (v1321) which has better error reporting and also handles the RETURN statements being warned about (and some other operators).

Thanks, Rob for such a swift response.

[Blackraeven]

Re certain problems with European locos.

I couldn't get my favorite European locos to run at all in OR until I made some minor changes to the eng file. It would seem OR doesn't like the metric system. I converted Boiler Pressure to "290 PSI" and Boiler Volume to

"850*(ft^3)". I'm no expert or purist in such matters but my locos now run according to my expectations. My heartfelt thanks to the folks at OR for doing such a splendid job and giving new life to my favorite sim. I've toyed with

Railworks and keep coming back to OR-MSTS!

Blackraeven

[cjakeman]

Blackraeven, on 23 January 2013 - 12:57 PM, said:

It would seem OR doesn't like the metric system.

It's not difficult to add some conversions from imperial to metric (and internally OR uses metric units). The issue is that MSTS does not treat all units equally or predictably so it's pretty much impossible for us to build a system which does the right thing and also replicates MSTS oddities.

There will be developments in this area fairly soon but, for now, it's possible to keep your unchanged metric MSTS eng files and also have a parallel set of imperial eng files just for OR.
You do this by adding an "\openrails\" folder alongside the MSTS eng file and putting your OR eng file into that new folder.

Hope that helps,