[Eldorado.Railroad]

I have been trying out the newer AI features.

The test is comprised of 1 player consist and 4 AI consists on any route for which an activity is created for. [These AI consists MUST start from a tile which the player trains DOES not reside on.]

The most current release I tested with was X2738 which results in a puzzling lack of particle emissions from AI units.

Thanks to the builds located here:
"http://james-ross.co.uk/projects/or/builds"

I tracked the working particle emitter for AI to release x2491, X2492 and beyond do not display particle emissions for AI units. At first I thought that maybe this was an error in an .eng file, but it appears that the code change made to ParticleEmitter.cs [twpol, X2492 Fix for emitter particles being frame-aligned, should improve random feel] is the culprit.

If anybody else would like to verify my results please post your findings here. The two files to download from "http://james-ross.co.uk/projects/or/builds" are:
Open Rails X2491 2014-09-15 19-59-12.zip (should work correctly for AI units)
Open Rails X2492 2014-09-15 21-19-11.zip (does NOT work correctly for AI units)

Thank you James for keeping those old builds available. They are indispensable for tracking down bugs.

************** Important Update **************
If your running AI train starts ON THE SAME tile that your player trains resides, particle/smoke emissions occur as expected. I only tested this particular scenario with ONE AI train.

[James Ross]

Eldorado.Railroad, on 27 December 2014 - 10:13 PM, said:

I tracked the working particle emitter for AI to release x2491, X2492 and beyond do not display particle emissions for AI units. At first I thought that maybe this was an error in an .eng file, but it appears that the code change made to ParticleEmitter.cs [twpol, X2492 Fix for emitter particles being frame-aligned, should improve random feel] is the culprit.

Could you compile (the latest code is fine) with DEBUG_EMITTER_INPUT enabled and see what it prints for the P= value for the AI emitter? Best thing is probably to only have one AI and remove all emitters from the player locomotive if you can. The P= value is the "particles per second" value which, if too low (below 0.1), might be at fault (given the changes in X2492).

[railguy]

AI diesel emissions are working fine for me in X2738. The loose consist smoking bug is still there, of course, but the AI units behave just like they are supposed to.

[Eldorado.Railroad]

railguy, on 28 December 2014 - 07:40 AM, said:

AI diesel emissions are working fine for me in X2738. The loose consist smoking bug is still there, of course, but the AI units behave just like they are supposed to.

Please specify what hardware and software you are using. AMD/ATI or Nvidia...etc.

When you say AI emissions are working fine in X2738, are you duplicating one player and 4 running AI consists that I mentioned in the OP?

Thanks!

[Eldorado.Railroad]

James Ross, on 28 December 2014 - 02:55 AM, said:

Could you compile (the latest code is fine) with DEBUG_EMITTER_INPUT enabled and see what it prints for the P= value for the AI emitter? Best thing is probably to only have one AI and remove all emitters from the player locomotive if you can. The P= value is the "particles per second" value which, if too low (below 0.1), might be at fault (given the changes in X2492).

At the moment I have no machine available to compile Open Rails. What I can do is arrange a test such that there are no emitters on the player locomotive and with only one running AI lcomotive in the activity see if its emitters are running....I will try this with one and two emitters on the AI locomotive.

Thank you!

[James Ross]

Eldorado.Railroad, on 28 December 2014 - 12:30 PM, said:

At the moment I have no machine available to compile Open Rails. What I can do is arrange a test such that there are no emitters on the player locomotive and with only one running AI lcomotive in the activity see if its emitters are running....I will try this with one and two emitters on the AI locomotive.

I don't think that will necessarily tell us anything new, without the debug logging, but you're welcome to try and report back anything interesting.

[Eldorado.Railroad]

James Ross, on 28 December 2014 - 12:53 PM, said:

I don't think that will necessarily tell us anything new, without the debug logging, but you're welcome to try and report back anything interesting.

Interesting, perhaps not, but in x2942 with the emitters removed from the player locomotive and only one emitter programmed on the AI locomotive, there are NO particle emissions to be seen. In X2941, things behave as expected, no emissions from the player locomotive, but one emitter is visible on the AI locomotive.

If this is of interest, at the start of the activity, the AI train is not visible because it is located and running several tiles away, I could find out just how many tiles if you would like.

If you could compile the code with the debug info mentioned above I will test and report back in a timely fashion. Just glancing at the code, it could be that the AI locomotive which is not in view, never gets "promoted" to being a possible emitter.

Thanks

[Csantucci]

Hi Eldorado,
please try replacing Runactivity.exe and RunactivityLAA.exe with the ones attached here and check if the problem is solved and no new problems have been introduced.
 Runactivities.zip (1.17MB)
Number of downloads: 190

[Csantucci]

Fixed in x2756.

[Eldorado.Railroad]

From 03 January 2015 - 12:00 PM:

Fixed in x2756.

Carlo,

I will try and test very carefully the new experimental release. Thank you very much for all your efforts in debugging and compiling/releasing the fix.

Steve

[Eldorado.Railroad]

Csantucci, on 03 January 2015 - 12:00 PM, said:

Fixed in x2756.

Tested extensively. Thank you for fixing this bug!

[railguy]

Maybe not directly related in the code, but the full-throttle loose consist smoking bug is still there. Oddly, the AI smoking bug that Eldorado noted (and has been fixed) never showed up in any of the dozen or so activities that I've run since the bug was identified. I do wonder if some of these issues are video card or other hardware/software-specific.

[Eldorado.Railroad]

railguy, on 05 January 2015 - 08:24 AM, said:

Maybe not directly related in the code, but the full-throttle loose consist smoking bug is still there. Oddly, the AI smoking bug that Eldorado noted (and has been fixed) never showed up in any of the dozen or so activities that I've run since the bug was identified. I do wonder if some of these issues are video card or other hardware/software-specific.

Would you please tell us what your OS/hardware/software setup is (detail is very important)? This might help.

Have you filed a formal bug complaint about the loose consist AI units being at (full throttle) for particle emissions?

Have you tried the various experimental releases to see when this problem started exactly?

Thanks for all you hard work and patience. Bug hunting is serious work.

[railguy]

Windows 7 Professional
3.4 gHz Intel Core i7-3770
64-bit operating system
16.0 gB RAM
AMD Radeon 7700 Video Card 1 gB Memory

The loose consist smoking problem has been around for well over a year and probably was what was reported as a bug in #1184649 back in 2013.

[Eldorado.Railroad]

railguy, on 06 January 2015 - 07:06 PM, said:

Windows 7 Professional
3.4 gHz Intel Core i7-3770
64-bit operating system
16.0 gB RAM
AMD Radeon 7700 Video Card 1 gB Memory

Thank for this info. As a footnote I also found the bug I mentioned in this post with and AMD/ATI card. I will have to see just how far back this problem started, which takes some time to track down.