PDA

View Full Version : internal test builds 77730 & 77738



WindWalkr
August 6th, 2015, 01:56 AM
Hi all,

77730 - Windows.
77738 - Mac OS X.

Some small fixes to validation (including a case where some newer LOD errors were not fully suppressed on older some assets.)
Some stability fixes.
Minor improvements to the asset preview stats.
More work aimed at the 3D text color issues.

chris

whitepass
August 6th, 2015, 06:56 PM
Quick test run and the Red/Green arrows on junctions are not showing and most(all?) roiling stock in CM view rotate around one end.

pcas1986
August 7th, 2015, 01:24 AM
Is there anything in particular you want us to test?

I'm not seeing anything different in Preview Asset and the dropdown lists still don't work in Windows (10).

JCitron
August 7th, 2015, 01:44 PM
Quick test run and the Red/Green arrows on junctions are not showing and most(all?) roiling stock in CM view rotate around one end.

I can confirm both issues.

The rolling stock rotating clock-wise on one end is weird but I can see how this occurred as it picks up on a reference point on the end instead of the center of the asset.

The switches can be an issue without the arrows if we're not using the procedural track so this definitely needs fixing.

John

Kevin16c
August 7th, 2015, 09:31 PM
I don't have the fastest internet here so I'm usually a bit behind getting the latest beta build, but after I finally got it installed today It seems that when testing it out it spits this error when downloading from the dls:

Is this intentional or a bug with the build?
- This Trainz version does not allow the installation of third-party content.

Also the Eucalypt Trees and the European Birch Speed trees among others (hit or miss with the rest of the trees) are still broken and have green and pink grids as the textures.

WindWalkr
August 7th, 2015, 11:25 PM
- This Trainz version does not allow the installation of third-party content.

Please check your settings and confirm that your MyTrainz account details are configured properly; perhaps try using "Start Trainz" to confirm this. Then try downloading; do you get the same issue?

chris

Kevin16c
August 8th, 2015, 12:52 AM
It seems that it sorted itself after I launched the game (start trainz). But I did notice some peculiar behavior with locomotive sounds. The whistle and bell are not fading out as you move away from the engine you are driving. For instance, placing the locomotive down at the corner of one base board and moving to the opposite corner of not the same tile but the opposite corner of an adjoining (effectively a two base-board gap between you and the locomotive) tile the bell and whistle/horn sounds can be heard as if you were next to the locomotive. Also the same can be said when you zoom the camera out to its maximum distance, not the distance before it turns into the "map view" but the maximum distance after it switches to the "map view" the bell and whistle sounds can still be heard as if you were standing right next to them instead of several miles away. They should fade away just like the ambient sounds from the scenery.

JCitron
August 8th, 2015, 09:19 AM
I spent most of my day in Surveyor replacing trees and modifying a very large route I've been working on for a few years now. In the T:ANE retail version, the program has crashed numerous times during the editing session. The good news is I did not have one CTD all day.

I will continue to testing...

John

pcas1986
August 11th, 2015, 03:38 AM
For Chris:

Just to reiterate what I asked in an earlier post - is there anything in particular you would like us to test? I'm still messing around with LOD and my current project but the amount of work means I haven't actually been in a position to test it.

WindWalkr
August 11th, 2015, 03:53 AM
Just to reiterate what I asked in an earlier post - is there anything in particular you would like us to test? I'm still messing around with LOD and my current project but the amount of work means I haven't actually been in a position to test it.

Not really. If you're happy, I'm happy. This isn't a beta testing process so I'm not asking you for any specific testing. If I want feedback on some change, I'll be specific about it. Otherwise, it's up to you guys to bring anything to my attention if you want something investigated.

cheers,

chris

andi06
August 11th, 2015, 04:51 AM
Don't have much time at the moment but general comments follow:



I seem to be getting roughly 10% better frame rates in this build, also slightly smoother but stability is much reduced. I had one serious blue screen crash that took half a dozen reboots to fully recover from. I understand that these random crashes may not be a serious issue for test builds but they seem to be a constant - have you been able to identify the cause(s) so that they can be dealt with at the appropriate time?




Procedural track components are OK now and I no longer have validation errors per-se. However I'm still upset that these LOD errors are now live in game with the proposed facility to suspend them for assets in development not yet implemented. Its a serious PITA to have to introduce LOD before I'm ready to do so. The continued absence of Trainzutil/compile is another similar irritant. I know what you are going to say on this so don't bother - I'm simply observing that issues like these are preventing serious asset development in TANE, the sooner they are available the better the flow of TANE assets.




Haven't noticed any change in Preview except that it is now reluctant to load some assets - I'm having intermittent crashes loading the Autocoach and I can't load <kuid:122285:700> at all.




Am I correct in assuming that the draw call/polycount values shown when an asset is zoomed out so far that it becomes invisible in the Preview window approximates to the minimum resource used by that asset when visible in game? If this is anything like correct then the Auran Duchess is consuming 75,31448 at 10 kilometres viewing distance. Whilst this might attract some flak in its own right it does provide a perspective on what E2 is having to deal with and why performance might be less than ideal. What might be involved in producing a small route using only TANE efficient assets to demonstrate what the game is actually capable of?

WindWalkr
August 11th, 2015, 08:09 AM
I had one serious blue screen crash that took half a dozen reboots to fully recover from. I understand that these random crashes may not be a serious issue for test builds but they seem to be a constant - have you been able to identify the cause(s) so that they can be dealt with at the appropriate time?

Blue Screens are basically driver faults. The best thing you can do there is generally to make sure that you have your OS-level crash reporting enabled so that the vendor can receive feedback and hopefully resolve the issues. Obviously if we have a reproducible case (ie. "do X then Y and receive a bluescreen") then we might be able to develop a workaround or involve the vendor directly but beyond that there's little that we can do to help.

Application crashes are potentially very different; these are quite likely our fault (although driver or hardware faults are still possible) and either a repro case (even a fairly involved one) or a crash dump can help resolve these. We fix these as we find them.




I'm still upset that these LOD errors are now live in game with the proposed facility to suspend them for assets in development not yet implemented. Its a serious PITA to have to introduce LOD before I'm ready to do so.

Noted.




The continued absence of Trainzutil/compile is another similar irritant.

I'm not sure what problem you're having with this one?




Haven't noticed any change in Preview except that it is now reluctant to load some assets - I'm having intermittent crashes loading the Autocoach and I can't load <kuid:122285:700> at all.

I didn't have any problems loading that asset into the preview window. Please retry in the next build.




Am I correct in assuming that the draw call/polycount values shown when an asset is zoomed out so far that it becomes invisible in the Preview window approximates to the minimum resource used by that asset when visible in game?

Yes, the preview window is in fact part of the game, not a separate utility. This is not emulation, this is the real thing. You can actually load up a route or session into the preview window if you really want.




If this is anything like correct then the Auran Duchess is consuming 75,31448 at 10 kilometres viewing distance.

I can't speak for that specific asset, but concerns over the LOD usage (or lack thereof, in some cases) has been raised. It is definitely accurate that "internally created" content is not immune to this sort of problem. Quotes because we don't actually create much content internally.




Whilst this might attract some flak in its own right it does provide a perspective on what E2 is having to deal with and why performance might be less than ideal.

It certainly doesn't help.




What might be involved in producing a small route using only TANE efficient assets to demonstrate what the game is actually capable of?

Depends on whether you're talking about efficiency or something else. The preview window makes it easy enough to demonstrate whether a given item of content is any good in most areas, performance-wise, so careful selection of content is really all that's needed here.

chris

pcas1986
August 11th, 2015, 08:28 AM
...


Haven't noticed any change in Preview except that it is now reluctant to load some assets - I'm having intermittent crashes loading the Autocoach and I can't load <kuid:122285:700> at all.

...


I didn't have any problems loading that asset into the preview window. Please retry in the next build.
chris

This loco loads in Preview Asset for me (77730).

andi06
August 11th, 2015, 08:55 AM
I'm not sure what problem you're having with this one?
Its simple really. I'm happy that installing scripted assets to test the script is a big improvement on TS12. If and when we get access to the TANE command line so that we can install a set of assets and then call up a test session with a single keypress it will be excellent. However installing an asset to check the script syntax (the only way that this can be done at the moment) is overkill. We need trainzutil/compile (even if we need to specify additional include paths) to get a quick syntax check.


I didn't have any problems loading that asset into the preview window. Please retry in the next build.
Its working now - perhaps I was dreaming.


I can't speak for that specific asset, but concerns over the LOD usage (or lack thereof, in some cases) has been raised. It is definitely accurate that "internally created" content is not immune to this sort of problem.
I'm not specifically gunning for your own assets but examples are helpful, it would be unfair to single out third party content and none of my own stuff (even where it doesn't have LOD at all) uses that many resources at LOD-0 let alone LOD-3.


Depends on whether you're talking about efficiency or something else.
I was thinking aloud along these lines:

1. You are getting a lot of flak about TANE performance, not everyone believes that content is the prime issue - quite a few are stating that the graphics engine is the real problem.
2. Its not enough to have one or two efficient assets while the route as a whole is still using 25K poly scenery items without LOD.
3. For you to prove that E2 is worth its salt you need to demonstrate what it can do when it isn't held back by poor content.
4. What might be involved in producing a small route using only TANE efficient assets to demonstrate what the game is actually capable of? ... and would it be worth while to consider commissioning something to include with the game?

WindWalkr
August 11th, 2015, 10:26 AM
We need trainzutil/compile (even if we need to specify additional include paths) to get a quick syntax check.

You have it, unless I'm horribly misunderstanding what you're talking about.



I'm not specifically gunning for your own assets but examples are helpful, it would be unfair to single out third party content and none of my own stuff (even where it doesn't have LOD at all) uses that many resources at LOD-0 let alone LOD-3.

Oh, I'm not worried about calling a spade a spade. We need to fix all the assets. Pretty much everyone's doing a terrible job and we all need to stop it. It's work, but it's not as hard as all that, as I'm sure you're very aware.




1. You are getting a lot of flak about TANE performance, not everyone believes that content is the prime issue - quite a few are stating that the graphics engine is the real problem.

There are lots of causes. Content is certainly a big one. Unrealistic expectations is another big one. The engine isn't fully mature in some areas, that's important if your route or hardware touches on one of those areas. None of these alone is to blame.



2. Its not enough to have one or two efficient assets while the route as a whole is still using 25K poly scenery items without LOD.

Agreed.



3. For you to prove that E2 is worth its salt you need to demonstrate what it can do when it isn't held back by poor content.
4. What might be involved in producing a small route using only TANE efficient assets to demonstrate what the game is actually capable of? ... and would it be worth while to consider commissioning something to include with the game?

I'm not the one making that call, but my take on this is: No.

There are plenty of routes out there that run just fine, by virtue of not being as taxing. Sure, they may not be as scenery-dense as some of the really heavy routes, but they serve as a good demonstration of what can be done if you keep things in check. I don't think that people who are having performance issues really want to be "proven wrong". They want a magic pill to make the performance issues go away. In the absence of any magic pills, we'll need to keep chipping away at the other things- including improving content (both directly, and by improving the tools which are used to improve the content), improving the code's ability to handle various special cases, and adding better custom support for some of the more less mainstream hardware configurations.

Content is an important part of that, which is why I've spent a fair bit of time here helping to explain where we're going wrong, and why we've been introducing tools and validation aimed at resolving this, but it's far from the only part of it.

chris

andi06
August 11th, 2015, 11:54 AM
You have it, unless I'm horribly misunderstanding what you're talking about.
So we do. Its just not listed in trainzutil help.

Are there any other helpful goodies hiding away?

TANE -nocrash ?
Trainzutil -reinstate-user-settings ?
TANE -wake-bergmann-up ?:)

ianwoodmore
August 11th, 2015, 03:58 PM
Updated to WIN 10 Pro Retail from WIN 8.1 Pro with Media Center, and installed TANE build 77730.
In maxi instal carried out manual database rebuild using previous TANE Userdata in conjunction build 77730 core.
After instal, I invoked a DBR which completed in 7.2 hours but resulted in extremely high number (thousands) of false errors in comparison to previous build.

eg Category "Fixed Track" showed 750 errors. After OFE and Commit only 47 were genuine.

Note majority of assets are now at TB 3.7 as I hunt down causes of non-compliance.

In addition, CM seems reluctant to update its displays no matter what default or custom filter is in focus.
Then it is necessary to change filter and back again.
Particularly prevalent when exploring faulty dependencies where dependency of dependency involved.

While this is a PITA at moment I can continue working, albeit at a slower pace, as everything has to be double checked rather than believe that CM is correct.
I'll revisit this problem at next build.

ianwoodmore
August 11th, 2015, 04:37 PM
@ZEC

Australian Modern Region kuid:103475:60030

When upversioned to TB 3.7 shows modified, Faulty dependencies

List dependencies

Car 51 and some others show default-night\night-mesh-base has value of "default". Set to "lod-0" clears the fault.

Car 53

! This asset uses an obsolete trainz-build number. Trainz-build numbers below 3.5 are no longer supported. (I left this one at TB 3.0)
- Alias mesh 'car53_lod1.im' not found in latest version of dependency <kuid:60850:29053> (<kuid2:60850:29053:2>)
- Alias mesh 'car53_lod2.im' not found in latest version of dependency <kuid:60850:29053> (<kuid2:60850:29053:2>)
! The meshes in LOD level 2 must total at least 20% fewer polygons than the next higher LOD: 0: 747, 1: 0, 2: 0

Car 54

- Alias mesh 'car54_lod1.im' not found in latest version of dependency <kuid:60850:29054> (<kuid2:60850:29054:2>)
- Alias mesh 'car54_lod2.im' not found in latest version of dependency <kuid:60850:29054> (<kuid2:60850:29054:2>)
! The meshes in LOD level 2 must total at least 20% fewer polygons than the next higher LOD: 0: 809, 1: 0, 2: 0

Toyota Yaris

- Alias mesh 'toyota yaris.im' not found in latest version of dependency <kuid:50587:27237> (<kuid2:50587:27237:1>)

extract from config:

kind "scenery"
kuid <kuid:103475:29011>
username "Toyota Yaris - Traffic - TS2010"
trainz-build 3.7
category-class "BH"
category-region "AU;GB;IE;UK;US"
category-era "1980s;1990s;2000s;2010s"


mesh-table
{
default
{
mesh-asset <kuid:50587:27237>
mesh "Toyota Yaris.im"
auto-create 1
}
}

mesh-asset points to:

kind "scenery"
kuid <kuid2:50587:27237:1>
username "Truck Landscaping"
trainz-build 3.7
category-class "VL"
category-region "00"


mesh-table
{
default
{
auto-create 1
mesh "Truck Landscaping.im"
}
}

ianwoodmore
August 11th, 2015, 05:08 PM
Author BUGOR kuid group 502415 A very prolific creator.

Of 3,739 assets some 2,047 of 2013/14 vintage are faulty.

These assets are predominantly used in these routes, session and buildings:

<kuid2:564595:1003:13> Uzlovaya v. 1.9 sU
<kuid:502415:100552> BGR Gryazi-Voronezhskie
<kuid2:502415:100522:1> BGR Michurinsk-Ural
<kuid2:502415:19877:2> BGR track building 9
<kuid2:706538:100004:7> Mosti-Balezino-Zavodskaya sU Final MP Final version
<kuid2:706538:100033:5> Mosti-Balezino-Zavodskaya sU Final MP GLLC v1.04
<kuid:706538:100003> Mst-Blzn-Sln v.4.04 SemiFinal MP with pult DSP
<kuid:528856:101097> 9 OCT TRÈS BELLE ROUTE
<kuid:381869:100092> Teplovoznaya doroga TS12 z7 BETA 1.0. RUSSIA MAP
<kuid:381869:100072> Тепловозная дорога TS12 z7 BETA-версия 1.0.
<kuid:381869:100090> Тепловозная дорога TS12 z7 BETA-версия 1.0.
<kuid2:645887:100002:2> Mosti-Balezino-Zavodskaya sU Final MP(ЦППК)
<kuid:705946:100002> GREGOVO MULTIPLAYER v. 1
<kuid:605125:100028> GREGOVO_MP_SU
<kuid:605125:100064> GREGOVO_MP_SU
<kuid2:734579:100054:1> Butolino - Vyazanki (disel) beta MP v0.1
<kuid:734579:100071> Butolino - Vyazanki (disel) beta MP v0.2
<kuid:734579:100036> Butolino - Vyazanki (disel) beta v1
<kuid2:734579:100083:1> Butolino - Vyazanki (disel) Final MP
<kuid:560053:100085> Мосты-Балезино-Заводская sU v.1.4 МП
<kuid2:750156:100011:2> Gregovo MP sU New v.2
<kuid2:750156:100029:2> Gregovo: Zolotaya Osen
<kuid:750156:100004> GREGOVO_MP_SU NEW
<kuid2:570553:100200:6> Mosti-Balezino-Zavodskaya sU Final MP
<kuid2:570553:100058:7> Mst-Blzn-Sln v.4.06.1 SemiFinal MP
<kuid2:748664:2015:1> Southern Ural Big Map v.1
<kuid:549351:100049> Belorechensk
<kuid:724946:100002> GREGOVO v1.1 SU
<kuid:132864:100074> Тестовый полигон
<kuid:500336:100348> Uzlovaya v. 1.5 sU MP

Of those I have investigated it appears that mesh-asset kuid2:502415:100532:4 BGR Switch Main Library which is massive (600MB+) has missing meshes.

This fault mode has reached almost pandemic problems in this build, but at a cursory inspection validation is correct.

Typical validation report

- Alias mesh '1-9/sk16/ballast/none.im' not found in latest version of dependency <kuid:502415:100532> (<kuid2:502415:100532:4>)
- Alias mesh '1-9/sk16/ballast/none.im' not found in latest version of dependency <kuid:502415:100532> (<kuid2:502415:100532:4>)
- Alias mesh '1-9/sk16/ballast/none.im' not found in latest version of dependency <kuid:502415:100532> (<kuid2:502415:100532:4>)
- Alias mesh '1-9/sk16/ballast/none.im' not found in latest version of dependency <kuid:502415:100532> (<kuid2:502415:100532:4>)

WindWalkr
August 11th, 2015, 07:08 PM
Author BUGOR kuid group 502415 A very prolific creator.

Of 3,739 assets some 2,047 of 2013/14 vintage are faulty.

Yes. It seems that one specific mesh-library that he uses for nearly everything is missing a mesh that is referenced by nearly everything. I've opened a dialog with him but have yet to provide him with full details.

chris

ianwoodmore
August 11th, 2015, 08:17 PM
Thanks Chris,

This morning I continued my search for false errors in validation after the manual DBR.
I found another ~3,500 in Category "fixed track". View errors and warnings cleared some, but majority needed OFE/Commit action.
It looks like Fixed Track may have the majority of the false errors.

RPearson
August 12th, 2015, 05:24 AM
The bug with the new SE property of track vertices appears to be fixed. I re-did my prevoius tests:
1) After re-entering the data the track shows a smooth transition in both directions about the vertex.
2) The maximum value that is accepted for 1st parameter is now 1000.0. - Thanks 100 was too small.
3) The 2nd parameter still has no (practicable) limit applied. Most regulations (worldwide) limit SE to 7 inches or less which works out to approx 0.22 radians or less for standard gauge track. While you can turn the track upside down and more with large values for params 1 and 2 I personally don't see a need to limit this unless it could cause other problems.
4) The mouse over help for the 1st parameter still provides ambigious and basically wrong guidance.

Degree of curvature (DOC) has more than 1 common definition. With railroad track curves it is normally related to a chord length of 100ft or 30m or some other length which has to be specified. Using any of the common definitions results in the wrong input value. The correct value to input can be obtained by dividing the bank angle in radians by the horizontal curvature (as 1/radius in meters) IIRC you stated pretty much the same in your description of this input. In practive this is equal to the bank angle times the radius of the curve (with angles in radians and lengths in meters). There are other ways of course such as bank angle in degrees divided by DOC (100 ft chord definition) times 30.48.

You can consider curvature as the change in central angle per change in arc length which in the calculus of limits as the change in length becomes infintessimally small has a limiting value of 1/radius. If we deal with fixed radius curves we can arrive at the correct value by just using the change in central angle per a unit change in the arc length (the units have to be consistent and the resulting angluar change is in radians). This is what Roland assumed because frankly degree of curvature is ambigious without indicating how it is measured. His assumption worked of course. I just assumed you guys screwed up and meant curvature to start with and reached the same result. Using a unit chord or arc length to determine the change in central angle of the arc is not the common definition of DOC. Certainly not in railroad track surveying. And clearly the bank angle is measured in radians not degrees as per the guidance.

#4 - I got carried away but maybe you could give us a better description such as bank angle per unit curvature or some such thing.

Bob Pearson

WindWalkr
August 12th, 2015, 07:29 AM
Thanks Bob,

I won't act on that immediately but I've put a note in the system to get it improved and added your thoughts.

cheers,

chris

pcas1986
August 12th, 2015, 09:48 AM
So we do. Its just not listed in trainzutil help.

...

I'm having problems getting the compile option to work. Either it says Tane isn't running or it says it says it's getting an improper response from Tane. The latter causes Tane to die.

I have a crashdump if N3V are interested.

pguy
August 17th, 2015, 10:46 PM
Hi Chris

I have succeeded in installing Mac build 77738 ( there is still an error in the document : the displayed name has been updated and is now correct, but the link behind has not been updated and still points to a non existent file. But I have pasted the correct name in my navigator and so succeeded in downloading and installing the build.).


But I have now a new problem. There are now a few missing dependencies. Some previous build builtin seems now no longer builtin and need to be downloaded from DLS. And when I request a download, the download starts but when installing I get this error message :


- This Trainz version does not allow the installation of third-party content.

and I still have the missing dependencies ... Any idea to overpass this problem ?

Regards.
Pierre.

WindWalkr
August 17th, 2015, 11:17 PM
I'm having problems getting the compile option to work. Either it says Tane isn't running or it says it says it's getting an improper response from Tane. The latter causes Tane to die.

I have a crashdump if N3V are interested.

Sorry, missed this for some reason. Please hold off until the next build, and if you're still getting crashes then email in a crashdump at that point.

chris

WindWalkr
August 17th, 2015, 11:19 PM
But I have now a new problem. There are now a few missing dependencies. Some previous build builtin seems now no longer builtin and need to be downloaded from DLS.

Not surprising. The content set isn't something that we've been working on and likely has a number of issues.



- This Trainz version does not allow the installation of third-party content.

We're actively tracking what causes this, but the simple workaround is to click the "start trainz" button.

chris

ianwoodmore
August 18th, 2015, 12:26 AM
Content Manager - Default Today filter broken?

Date modified and Date installed do not reflect actual today date.

You can attempt to update this using the calender function and this works until next time you need to use Today filter.
Then filter shows last date on which you used it.

Tried to reset filter using the Today box in calender and then save but still not tracking date next time you use.

PC is latest WIN 10 update OS.

pcas1986
August 18th, 2015, 03:02 AM
Just had a crash in 77730 while testing several assets for someone who wanted to know if a particular feature worked or not. Was running a session and just coupled two locos together and then got a warning from Windows (10) saying program had stopped working and did I want to debug. I said no. There was no crash dump. The locos concerned use Andi's Superscript although there was nothing to suggest that was the issue.

Do you want me to try and reproduce this? This may be difficult since I'd been trying a number of different versions of this particular loco.

andi06
August 18th, 2015, 04:09 AM
The locos concerned use Andi's Superscript although there was nothing to suggest that was the issue.
But you thought you might just as well implicate me anyway? :)

pcas1986
August 18th, 2015, 04:36 AM
But you thought you might just as well implicate me anyway? :) Not deliberately, but I vaguely recall you saying there were problems with Superscript within T:ANE but I could be wrong - I frequently am.

The traincar assets in question popup a browser when they are coupled together. This surprised me as I don't recall seeing it before. I'll admit I don't use Superscript, not because I don't think it a wonderful asset (tic), but I'm silly enough to think I can write my own scripts.

I was asked by Mick Berg to test ACS in conjunction with the animated buffers he uses on his battery locos (London Underground). Mick's battery locos use Superscript. After coupling one of his battery locos with his 1938 TubeStock carriages, the train started moving and then T:ANE crashed. I don't get many T:ANE crashes but then I don't run sessions very often. Most of my time is counting polys and wondering why 2+2 = 10. :o

andi06
August 18th, 2015, 05:13 AM
The main issue with Superscript and TANE relates to the fact that the included coupler meshes are *.im and tend to fall foul of the LOD rivet counting validation. I've modified my own copy of the library to include *.lm.txt versions but I haven't uploaded it yet (in case something else turns up)


The traincar assets in question popup a browser when they are coupled together.
If the browser has 'Superscript' in the title bar this is a debug mode browser window. To suppress it change 'debug-122285' to 0 in your copy of the Superscript library asset.

I believe that Mick Berg used a custom ACS message, the browser should help if it is showing the ACS messages being received. This could also be a TANE animation issue.


Most of my time is counting polys and wondering why 2+2 = 10. :o
I'm pretty sure that I've got to the bottom of this. See this post. (http://forums.auran.com/trainz/showthread.php?122290-LOD-Accounting&p=1433446#post1433446)

pcas1986
August 18th, 2015, 06:46 AM
....

If the browser has 'Superscript' in the title bar this is a debug mode browser window.
It does.


...To suppress it change 'debug-122285' to 0 in your copy of the Superscript library asset.

There was no such line but I added it anyway. But the asset promptly got broken due to allegedly missing texture.txt files. So I reverted it. It left me wondering why the asset would default to debug mode. This is the built-n version (20).


...

I believe that Mick Berg used a custom ACS message, the browser should help if it is showing the ACS messages being received. This could also be a TANE animation issue...
It's certainly punching out a lot of messages MessageBox style. ACS appeared to be working prior to the crash so I didn't think it relevant. After my earlier post I ran the same session with coupling and uncoupling a couple of times but T:ANE didn't crash.


...I'm pretty sure that I've got to the bottom of this. See this post. (http://forums.auran.com/trainz/showthread.php?122290-LOD-Accounting&p=1433446#post1433446)
I saw that and was thinking of walking back through my current project with the same technique but the asset might be too complex. Maybe I'll try and find a simpler asset.

andi06
August 18th, 2015, 07:19 AM
There was no such line but I added it anyway. But the asset promptly got broken due to allegedly missing texture.txt files. So I reverted it. It left me wondering why the asset would default to debug mode. This is the built-n version (20).

'extensions/debug-122285 1' can be added either in the library itself or in the client asset. By default it should be absent (and hence zero) in both and this will certainly suppress the browser. I will need to check the code to see what combination will enable it.

pcas1986
August 18th, 2015, 07:39 AM
'extensions/debug-122285 1' can be added either in the library itself or in the client asset. By default it should be absent (and hence zero) in both and this will certainly suppress the browser. I will need to check the code to see what combination will enable it.

Sorry, my mistake. I was testing different versions of basically the same asset in the same session. One of those assets had "debug-122285 1" in its extensions container but not the asset I was examining. I'll talk to Mick about it.

Pencil42
August 20th, 2015, 11:17 AM
It appears that m.gloss isn't reflecting properly in this build (77730)?

Build 77730:
http://www.carsoncarshops.com/images/web/bad_m_gloss.jpg


Build 76401:
http://www.carsoncarshops.com/images/web/good_m_gloss.jpg

Curtis