Update TransX development

[Enjo]

Blixel:
Unfortunately the problem persists. In one of my scenarios BTC MFD hangs Orbiter, but is seems that it's not that often... I'll try to recreate it in a debug version and hopefully it will be something obvious, so that I don't need to redesign the library.

---------- Post added 01-16-14 at 12:17 PM ---------- Previous post was 01-15-14 at 07:47 PM ----------

Blixel:
Could you test it the way I described anyway? We'll see if there's any difference. If it persists at your end, I'll ask for testers in an another thread.

 

[blixel]

Blixel:
I might have found the solution... Please do the following:
1) By retrying the to load the scenario, try to bring your machine to the state where it hangs in a repeatable way.
2) Open Orbiter_NG.cfg and from the active modules list remove the line:
BurnTimeCalcMFD
3) Test if it still hangs on reloading of the scenario.

I'm having a hard time getting back to the point where it hangs in a repeatable way. This morning, I loaded the scenario 11 times in a row. The first 10 times it loaded without hanging, the 11th time it hung. After Orbiter shutdown and restarted, the scenario loaded again even without removing BurnTimeCalc. Interestingly enough. The scenario loaded 10 times in a row again but hung on the 11th time.

I tried loading the Quickstart scenario 11 times in a row to see if it was just a general problem with my setup, but I was able to load the Quickstart scenario 13 times in a row without Orbiter hanging.

On a related note, last night I was doing a bunch of test flights in the standard DG where I took off from KSC and got to orbit. Orbiter was consistently crashing on me if I had BTC loaded. When I would get up to about 5,500 m/s (can't remember the exact Sim time), and I would press F8, or if I would try to bring up a different MFD, Orbiter would crash. If I did NOT have BTC running, it wouldn't crash.

Anyway, with regards to this scenario, after the 11th time loading Orbiter (when it hangs), here is the Orbiter.log on my machine:

**** Creating simulation session
D3D9Client: [DirectX 9 Initialized]
D3D9Client: Sytem has XNA math support
D3D9Client: [3DDevice Initialized]
D3D9Client: [Compiling Effects for Shader Model 3.0]
D3D9Client: [Loading Stars]
D3D9Client: [Loading Constellations]
D3D9Client: [D3D9Client Initialized]
Joystick throttle: Z-AXIS
Joystick throttle control detected
Module Sun.dll ............... [Build 100830, API 100830]
VSOP87(E) Sun: Precision 1e-006, Terms 554/6634
Module Mercury.dll ........... [Build 100830, API 100830]
VSOP87(B) Mercury: Precision 1e-005, Terms 167/7123
Module Venus.dll ............. [Build 100830, API 100830]
Module VenusAtm2006.dll ...... [Build 100830, API 100830]
VSOP87(B) Venus: Precision 1e-005, Terms 79/1710
Module Earth.dll ............. [Build 100830, API 100830]
Module EarthAtmJ71G.dll ...... [Build 100830, API 100830]
VSOP87(B) Earth: Precision 1e-008, Terms 2564/2564
Module Moon.dll .............. [Build 100830, API 100830]
ELP82: Precision 1e-005, Terms 116/829
Module Mars.dll .............. [Build 100830, API 100830]
Module MarsAtm2006.dll ....... [Build 100830, API 100830]
VSOP87(B) Mars: Precision 1e-005, Terms 405/6400
Module Phobos.dll ............ [Build ******, API 060425]
Module Deimos.dll ............ [Build ******, API 060425]
Module Galsat.dll ............ [Build 100217, API 100215]
Module Jupiter.dll ........... [Build 100830, API 100830]
VSOP87(B) Jupiter: Precision 1e-006, Terms 1624/3625
Module Io.dll ................ [Build 100217, API 100215]
Module Europa.dll ............ [Build 100217, API 100215]
Module Ganymede.dll .......... [Build 100217, API 100215]
Module Callisto.dll .......... [Build 100217, API 100215]
Module Satsat.dll ............ [Build 100215, API 100212]
Module Saturn.dll ............ [Build 100830, API 100830]
VSOP87(B) Saturn: Precision 1e-006, Terms 2904/6365
Module Mimas.dll ............. [Build 100215, API 100212]
SATSAT Mimas: Terms 113
Module Enceladus.dll ......... [Build 100215, API 100212]
SATSAT Enceladus: Terms 33
Module Tethys.dll ............ [Build 100215, API 100212]
SATSAT Tethys: Terms 101
Module Dione.dll ............. [Build 100215, API 100212]
SATSAT Dione: Terms 59
Module Rhea.dll .............. [Build 100215, API 100212]
SATSAT Rhea: Terms 68
Module Titan.dll ............. [Build 100215, API 100212]
SATSAT Titan: Terms 100
Module Iapetus.dll ........... [Build 100215, API 100212]
SATSAT Iapetus: Terms 605
Module Uranus.dll ............ [Build 100830, API 100830]
VSOP87(B) Uranus: Precision 1e-006, Terms 1827/5269
Module Miranda.dll ........... [Build ******, API 060425]
Module Ariel.dll ............. [Build ******, API 060425]
Module Umbriel.dll ........... [Build ******, API 060425]
Module Titania.dll ........... [Build ******, API 060425]
Module Oberon.dll ............ [Build ******, API 060425]
Module Neptune.dll ........... [Build 100830, API 100830]
VSOP87(B) Neptune: Precision 1e-006, Terms 391/2024
Finished initialising world
Module DeltaGlider.dll ....... [Build 131111, API 100830]
Module LuaInline.dll ......... [Build 100830, API 100830]
Finished initialising status
Finished initialising camera
Finished initialising panels
Finished setting up render state
D3D9Client: [Scene Initialized]

 

[Enjo]

There's nothing suspicious in this log... except that I can't find an entry saying that TransX.dll was loaded.

Regarding BTC's CTD:
You were dealing with an original bug, that I discovered myself just yesterday. A fix is attached.

 

[Enjo]

Let me stress that we MUST find out whether it's the ModuleMessaging library fault before doing anything. You said that on the previous version it didn't hang, but what you've written recently makes the previous test not so convincing anymore, wouldn't you agree? During the weekend I will perform my own tests as well, to try to verify the source (or at least lack of) further.

 

[blixel]

Let me stress that we MUST find out whether it's the ModuleMessaging library fault before doing anything. You said that on the previous version it didn't hang, but what you've written recently makes the previous test not so convincing anymore, wouldn't you agree? During the weekend I will perform my own tests as well, to try to verify the source (or at least lack of) further.

I agree. The behavior is quite strange. I can say this though. I sent the scenario to dgatsoulis and he said it hung on his system also. I think he's too busy with work at the moment to keep up with this thread, but I'm sure he'll share his findings when he gets a chance.

Hopefully between the 3 of us, we can nail down the exact problem.

 

[Ripley]

If you need it, I can test that scenario on my XP with inline client (which I forgot to do before just because I...never use it anymore!).
That will happen NET tonight (UTC+1).

Sorry it took longer than expected.

I recreated the exact same Launchpad configuration as my orbiter_ng.exe.
The same scenario doesn't hang in XP & inline client either.
Repeated executions of same scenario, some with time accel and some without but NO CTDs at all.

 

[Enjo]

Thanks very much for a comprehensive test, delayed or not.
So there's still not enough to convince me that it's my addons' fault, nor the MM SDK's, but a definitive confirmation I would get if somebody reported the same hang in DX9, but without using my latest addons...

Anyway, keep in mind, that if it really is the MM's fault, then I definitely cannot say that we can't roll back and not use this optional library at all. It's a matter of deleting just a few lines of code, no matter what further functionalities are added from now on. I'd really like to avoid deleting it though.

 

[Enjo]

For anybody whose Orbiter was hanging after installing Module Messaging library - please try this module pack and try to recreate the hanging situation.

 

[blixel]

For anybody whose Orbiter was hanging after installing Module Messaging library - please try this module pack and try to recreate the hanging situation.

I tried loading several different scenarios that were known to be problematic before. I tried loading them multiple times. With these dll's, Orbiter is not hanging.

I'm a bit hesitant to say that it's fixed though since the problem tended to be intermittent. (And since it didn't effect everyone. Ripley didn't seem to have any issue with how things were before. Perhaps a mysterious advantage that he has in running WinXP?) However, whatever you changed seems to be a step in the right direction, and possibly has solved the problem.

Input from more people is needed.

 

[Ripley]

I'll try making some FOI members betatest this freezing scenario issue. I already wrote there about TransX development.

---------- Post added 25-01-14 at 10:11 ---------- Previous post was 24-01-14 at 22:01 ----------

So for this betatest to succeed users need to install
- latest TransX
- latest BurnTimeCalcMFD
- latest LaunchMFD
- initial ModuleMessaging

And try with both inline and graphics client.
If user freezes (or CTDs), try the MM test library.
Right?

Are there redundant steps?

 

[Enjo]

I'll be right with you there. I need to finish some thing here.

---------- Post added at 11:45 AM ---------- Previous post was at 10:36 AM ----------

I'll try making some FOI members betatest this freezing scenario issue. I already wrote there about TransX development.

So for this betatest to succeed users need to install
- latest TransX
- latest BurnTimeCalcMFD
- latest LaunchMFD
- initial ModuleMessaging

And try with both inline and graphics client.
If user freezes (or CTDs), try the MM test library.
Right?

Are there redundant steps?

First of all, thanks for the offer. Here's how I imagine a proper test:
- get the latest files from the list above from OH
- try to recreate the hang bug in DX9 (details at the bottom of the post)
- try to recreate in with the inline client - the assumption is that it won't hang
- once the hang is reproduced in DX9, try to replace the DLLs (using JSGME) with the test releases that I will be posting in this thread and try to recreate
- please make sure whether the bug is gone or not, by making enough tests

The latest test release that I've attached recently has the MM lib completely disabled. Blixel reported that it doesn't hang, but guess what? It still hangs on my PC on my test scenario. Since I believe that the hang is not connected with MM (more tests please...), I reverted the changes to reinclude the MM lib, and made a debug release, so that I could break the simulation in debugger once it hangs, to see where the cursor is. The problem is that it never hangs on the debug version, which leads me to a suspicion that the problem is a result of a compiler bug during code optimizations. After all, it's an old compiler. A remedy to this would be compiling the addons with the latest MSVS, but this means that I couldn't debug the addons so easy as now, but it would still be possible. Also, since you, Ripley, reported no problems on DX7, and I got the same results as you, and BTC still uses the old drawing technique (from Orbiter 2006 API), another suspicion is that the bug is related to DX9's slight incompatibility with the old drawing technique.

Therefore my plan is the following:
1) Recompile the module pack in MSVS2005 using different compiler flags
2) Recompile under MSVS2013 Express
3) Reprogram BTC to make it use the latest drawing methods
4) Repeat 1) and 2)
For each of the point I will be releasing a module pack, so that others can operate under some kind of "defined states"

Here come the bug reproduction scenarios and infos:

Blixel's double TransX scenario:
[ame="http://www.youtube.com/watch?v=-P_D9uGv6gw"]Orbiter 2010 - Testing BTC 2.6 - YouTube[/ame]
Scenario post and further infos

My scenario. As the description says, my PC hangs 1/15 times, when I enable BTC in the left display.

BEGIN_DESC
Enable BTC MFD as left display - in some cases may hang Orbiter
END_DESC

BEGIN_ENVIRONMENT
  System Sol
  Date MJD 52033.8159066207
END_ENVIRONMENT

BEGIN_FOCUS
  Ship GL-01
END_FOCUS

BEGIN_CAMERA
  TARGET GL-01
  MODE Cockpit
  FOV 60.00
END_CAMERA

BEGIN_HUD
  TYPE Surface
END_HUD

BEGIN_MFD Right
  TYPE User
  MODE TransX
  Ship  GL-01
  FNumber 2
  Int 1
  Orbit True
  Vector  -1690776.91031 972224.238848 6295703.16702
  Vector  -7227.44699454 1818.35350713 -2222.6231407
  Double  3.98600439969e+014
  Double  52033.8159061
  Handle Earth
  Handle NULL
  Handle NULL
Select Target
 0 Escape
Autoplan
0 0
Plan type
0 0
Plan
0 1
Plan
0 0
Plan
0 0
Select Minor
 0 None
Manoeuvre mode
0 0
Auto-Center™
0 0
Base Orbit
0 0
Prograde vel.
 1  0
Man. date
 1  52033.8159057
Outward vel.
 1  0
Ch. plane vel.
 1  0
Intercept with
0 0
Orbits to Icept
0 0
Graph projection
0 0
Scale to view
0 0
Advanced
0 0
Pe Distance
 1  6633767.34509
Ej Orientation
 1  0.13962634016
Equatorial view
0 0
Finvars
  Finish BaseFunction
  Int 2
  Orbit False
  Handle Sun
  Handle Earth
  Handle Mars
Select Target
 0 Mars
Autoplan
0 0
Plan type
0 2
Plan
0 0
Plan
0 0
Plan
0 1
Select Minor
 0 None
Manoeuvre mode
0 0
Auto-Center™
0 0
Base Orbit
0 1
Prograde vel.
 1  0
Man. date
 1  52033.8157594
Outward vel.
 1  0
Ch. plane vel.
 1  0
Intercept with
0 0
Orbits to Icept
0 0
Graph projection
0 0
Scale to view
0 0
Advanced
0 0
Prograde vel.
 1  3022.61012323
Eject date
 1  52033.8420695
Outward vel.
 1  0
Ch. plane vel.
 1  751.349508762
Finvars
  Finish BaseFunction
END_MFD

BEGIN_SHIPS
ISS:ProjectAlpha_ISS
  STATUS Orbiting Earth
  RPOS 224769.14 -6491435.39 1788600.67
  RVEL -7555.705 125.361 1428.667
  AROT 30.00 0.00 50.00
  AFCMODE 7
  IDS 0:588 100 1:586 100 2:584 100 3:582 100 4:580 100
  NAVFREQ 0 0
  XPDR 466
END
Mir:Mir
  STATUS Orbiting Earth
  RPOS -6315650.88 -131844.93 -2155674.15
  RVEL 2501.108 -446.217 -7295.861
  AROT 0.00 -45.00 90.00
  AFCMODE 7
  IDS 0:540 100 1:542 100 2:544 100
  XPDR 482
END
Luna-OB1:Wheel
  STATUS Orbiting Moon
  RPOS -1939794.81 -1115682.25 -131.00
  RVEL 737.979 -1283.171 -0.097
  AROT 0.00 -0.00 -95.44
  VROT 0.00 0.00 10.00
  AFCMODE 7
  IDS 0:560 100 1:564 100
  XPDR 494
END
GL-01:DeltaGlider
  STATUS Orbiting Earth
  RPOS -1691078.66 972300.16 6295610.36
  RVEL -7227.349 1818.297 -2222.989
  AROT 138.02 69.10 131.48
  RCSMODE 2
  AFCMODE 7
  PRPLEVEL 0:0.572342 1:0.966258
  NAVFREQ 402 94 0 0
  XPDR 0
  AAP 0:0 0:0 0:0
END
SH-02:ShuttleA
  STATUS Landed Earth
  BASE Cape Canaveral:5
  POS -80.6745292 28.5197208
  HEADING 100.00
  AFCMODE 7
  PRPLEVEL 0:1.000000 1:1.000000
  NAVFREQ 0 0
  XPDR 0
  PODANGLE 0.0000 0.0000
  DOCKSTATE 0 0.0000
  AIRLOCK 0 0.0000
  GEAR 0 0.0000
  PAYLOAD MASS 0.0 0
END
LaunchMFD-Probe-GL-01:LaunchMFD-Probe
  STATUS Orbiting Earth
  RPOS -879898.76 -307299.55 -6945876.42
  RVEL 7211.179 -2049.023 -822.875
  AROT 0.00 -0.00 0.00
  AFCMODE 7
  NAVFREQ 0 0
END
END_SHIPS

BEGIN_ExtMFD
END

Thanks again.

 

[Ripley]

...First of all, thanks for the offer. Here's how I imagine a proper test:
- get the latest files from the list above from OH
- try to recreate the hang bug in DX9 (details at the bottom of the post)
- try to recreate in with the inline client - the assumption is that it won't hang
- once the hang is reproduced in DX9, try to replace the DLLs (using JSGME) with the test releases that I will be posting in this thread and try to recreate
- please make sure whether the bug is gone or not, by making enough tests
....

Ok, so I will hopefully start gathering some data with the test we both described, using blixel's original scenario.

I'll leave this bold quote out of the betatest for now, as I understand there will be come newer dlls from you.

...
If user freezes (or CTDs), try the MM test library.
...

Thread is located here

I didn't quite understand what I should do with the scenario you posted?
Should we test both of them, or just one?

 

[Enjo]

I didn't quite understand what I should do with the scenario you posted?
Should we test both of them, or just one?

Just open BTC in left display (MFD), and it should immediately hang on this event in one of the simulation sessions.
I'd ask to test both scenarios. For example, so far I wasn't able to reproduce the bug on Blixel's scenario, also due to lack of time recently. The more info we get, the better.

I'm downloading MSVC 2013 Express right now.

 

[Enjo]

Here are my planned test releases. They all contain MM SDK.

02 - compiled under VS 2013
03 - made the BTC inherit from MFD2 (new drawing technique), and compiled under VS 2005
04 - MFD2 as above, but compiled under VS 2013
05 - Optimizations disabled, VS 2005
06 - Optimizations disabled, VS 2013

As I said before, the best way to keep track of the tests, is to use JSGME for each test package.
To test the VS 2013 versions you will need the VC++ 2013 redistributables

 

[Ripley]

...I'd ask to test both scenarios...

FOI thread edited to reflect your wish.
:hail:

---------- Post added 26-01-14 at 15:51 ---------- Previous post was 25-01-14 at 19:59 ----------

It might not have anything to do with this, but I've just noticed that on blixel's video splash screen, he's using the "ultra-latest" D3D9 R12a experimental build.

 

[Enjo]

Sharp eye there!
I believe that the bug has much to do with DX9, as both you and me haven't had the problem in DX7 client (however input from more users would be welcome). It's just that I don't want to give up and blame DX9, but want to provide as much of compatibility as possible. It's annoying for me, as things don't work as advertised, but that's reality. However the fact that DX9 even supported the old MFD(1) implementation, which is used by the currently distributed version of BTC, and not only MFD2 based, is also a gift from Jarmonik, who was trying to provide compatibility in DX9 for older addons.

 

[Ripley]

Enjo, what kind of hardware/software info do you need, along with the test results?

...I believe that the bug has much to do with DX9, as both you and me haven't had the problem in DX7 client...

Wait.
In my initial batch of tests, I had one freeze on blixel's scn, with inline client, but the pc where I'm testing today is not my own pc.
Anyway, it is still XP, and with a vanilla Orbiter install for this purpose.
Also just updated DirectX to the latest (it's definitely not a gamer's pc).

 

[Enjo]

Enjo, what kind of hardware/software info do you need, along with the test results?

Let's say that it would be enough if enough people did the full test procedure on enough samples, telling only on which client and which test package it hung and on which not. Maybe let's add OS.

In my initial batch of tests, I had one freeze on blixel's scn, with inline client

Then I'm out of clue. But have you never had hangs without TransX? (I'm suggesting that it may be TransX-unrelated)

 

[Ripley]

I forgot to add that in my only freeze so far, Orbiter's sim time was 1 and not 0 (as it did with blixel).

Generally my Orbiter never crashes (nor freezes), not in a repeatable way anyway.
There may be an occasional random CTD here and there, but it's very seldom, nearly physiological.

---------- Post added at 18:43 ---------- Previous post was at 18:01 ----------

Can we use this syntax, to save some click-time?

orbiter.exe -s "Scenario blixel"

or

Modules\Server\orbiter.exe -s "Scenario enjo"

as recently described here?
http://orbiter-forum.com/showthread.php?p=451323&postcount=13

---------- Post added 27-01-14 at 00:38 ---------- Previous post was 26-01-14 at 18:43 ----------

Results from today's test:

1 freeze - blixel's scenario - inline - Enjo's default MFDs
2 freeze - blixel's scenario - D3D9 - Test-MM-04-MFD2-VS2013
1 freeze - blixel's scenario - inline - Test-MM-06-MFD2-NoOpt-VS2013
5 freeze - blixel's scenario - D3D9 - Test-MM-06-MFD2-NoOpt-VS2013

XP SP3, latest DirectX, 10 runs each, last set of DLLs looking particularly problematic (but, again, it was not a gamer's rig):

- Orbiter2010-P1 vanilla (NO OrbiterSound, NO Vinka, etc...)
- D3D9 R12 (+ symlinks)

From Orbit Hangar:
- TransX 2014.01.11 ModuleMessaging
- BurnTimeCalcMFD (BTC) 2.6
- Launch MFD - v. 1.6.2 for Orbiter 2010
- HUDDrawer SDK v.0.3
- ModuleMessaging SDK v. 1.0

 

[blixel]

I'm running these tests on my 2nd computer. The computer was formatted just a couple of weeks ago, so it has a brand new installation of Win 7 Home Premium x64 installed on it.

All tests are with Orbiter 2010 P1, OrbiterSound 4.0, D3D9Client R12a Experimental Build [January 15 2014]

Test-MM-02-VS2013 - compiled under VS 2013

I can't test this one because I get the following missing dll's. "This program can't start because MSVCP120.dll is missing from your computer." and "This program can't start because VesselHooking.dll is missing from your computer."

I didn't notice your note about the VC++ 2013 redistributables requirement before I did my testing. :facepalm: So I'll have to try these another time.

Test-MM-03-MFD2-VS2005 - made the BTC inherit from MFD2 (new drawing technique), and compiled under VS 2005

I am able to launch the "blixel" scenario 2 times in a row without it hanging. I don't really have time to try it 5 or 10 times.

Test-MM-04-MFD2-VS2013 - MFD2 as above, but compiled under VS 2013

Couldn't test at the moment.

Test-MM-05-MFD2-NoOpt-VS2005 - Optimizations disabled, VS 2005

I am able to launch the "blixel" scenario 2 times in a row without it hanging.

06 - Optimizations disabled, VS 2013

Couldn't test at the moment.