TimeForLuaClosing

[Scotfleiger]

Users are reporting problems with the LINDA add-on not closing when closing P3D 3.4.9 with FSUIPC4 4.9.5.7b. Having just done this update myself I am having the same problem. I note in the FSUIPC4 release notes that a new option TimeForLuaClosing has been added with a specific mention of LINDA. I have tied setting this value to 10 and 20 but neither LINDA not AS16 are closing when P3D is exited. With P3D 2.3 and previous version of FSUIPC4 the following FSUIPC4.ini statement correctly closed the add-ons:

[Programs]
RunIf1=KILL,C:\Sim Support\AS16_P3D\AS16.exe
RunIf4=KILL,C:\Sim Support\VAInterface Lite\VAInterfaceLite.exe
RunIf2=CLOSE,C:\P3D\Modules\linda.exe

Any suggestions on what the problem can be?

Note: the release notes (para 18) refer to TimForLuaClosing not TimeForLuaClosing (as appears in the .ini file).

[Pete Dowson]

This has been reported in another thread, and has been shown to be due to P3D not actually closing, but hanging or crashing after the attempt to close and before FSUIPC has had a change to action the KILL or CLOSE requests. If you check your Windows Event Viewer you may well find the entries for these occurrences.

It seems that something LINDA is doing, or possibly has instigated through FSUIPC, is causing problems with the recent releases of P3D. I think the LINDA folks will need to look into this. It is all an unknown to me, unless they can of course manage to narrow it down to a specific request.

Pete

 

[Scotfleiger]

Thank you Pete. As the 'LINDA person' I will investigate further. Is it just LINDA that is having these issues with no closing correctly?

[Pete Dowson]

2 minutes ago, Scotfleiger said:

Thank you Pete. As the 'LINDA person' I will investigate further. Is it just LINDA that is having these issues with no closing correctly?

Oh, hello Mr. LINDA! ;-)

The only two reports so far have been with LINDA. When LINDA is removed it is okay. Now it may be related to something in FSUIPC + P3D3.4, but I it would need narrowing down. If there's a way of doing a process of elimination with the multiple facets of the LINDA maze then perhaps this could be done.

BTW the other report is not far from here:
http://forum.simflight.com/topic/82221-runif1kill-seems-not-to-work-properly-after-update-to-4957b/

Pete

 

[Scotfleiger]

Hi Pete

The issue appears to be associated with the running of Linda.lua. If I place ipcready.lua and Linda.lua in a temporary folder and remove the [LUAFILES] entries, I can run P3D v3.4.9 and auto start Linda.exe and as16.exe. On closing P3D, both programs shutdown correctly about 10-15 secs later. However, when ipcready.lua and Linda.lua are restored to the /modules folder, P3D crashes on exit.

The crash sequence as reported by Event Viewer is .NET runtime error, followed by an application error in ntldll.dll and then APPCRASH for prepar3d.exe.

I was seeing P3D v3.3 crashing on occasions but always got an on-screen crash report. I am not seeing this with P3D v3.4. LINDA is distributed with and uses LUA5.1. Could this the cause?

[Pete Dowson]

5 hours ago, Scotfleiger said:

The issue appears to be associated with the running of Linda.lua. If I place ipcready.lua and Linda.lua in a temporary folder and remove the [LUAFILES] entries, I can run P3D v3.4.9 and auto start Linda.exe and as16.exe. On closing P3D, both programs shutdown correctly about 10-15 secs later. However, when ipcready.lua and Linda.lua are restored to the /modules folder, P3D crashes on exit.

Yes, that's what I would have assumed in any case. Obviously without any part of Linda operating inside P3D it is very unlikely indeed to be able to cause P3D to crash. But isn't "Linda.lua" part of LINDA? 

There are a great many Lua plug-ins running which don't have any problem with P3D closing.  I myself use many in relation to my cockpit. I would assume the process of elimination I mentioned would be oriented around the things you are asking Linda.lua to do.

BTW is there no Windows crash data available for this incident, from the Event Viewer?

5 hours ago, Scotfleiger said:

I was seeing P3D v3.3 crashing on occasions but always got an on-screen crash report. I am not seeing this with P3D v3.4. LINDA is distributed with and uses LUA5.1. Could this the cause?

Lua 5.1 is also the version FSUIPC includes. I don't know Linda at all, but I assume you need an external Lua interpreter too for some reason? Why do you think anything external to P3D could be a cause?

Pete

 

 

[Scotfleiger]

Hi Pete,

Thanks for the reply. For your information, LINDA comprises 2 elements: the Linda.lua handles the majority of the configuration and device handling and Linda.exe provides the UI for device assignment programming, handles the VRI devices and communicates with FSUIPC4.

There are no LUA errors being reported in the Event Viewer. The first fault indication is the .NET error on closing P3D and LINDA does not use this framework.

I will delve further.

[Pete Dowson]

49 minutes ago, Scotfleiger said:

There are no LUA errors being reported in the Event Viewer. The first fault indication is the .NET error on closing P3D and LINDA does not use this framework.

A .NET error with Prepar3D.exe as the process name? That's very strange.

LUA errors cannot show up in any Windows crash report for P3D unless a LUA DLL was actually running inside P3D, which would certainly not happen with the FSUIPC implementation as the Lua interpreter is embedded in FSUIPC itself.

FSUIPC does not use .NET either, and as far as I know neither does P3D -- at least it never installs anything for it, and .NET applications are pretty much always version specific, or as I thought. So if Windows reports a .NET error in P3D there must surely be some other add odd using .NET inside P3D. Ugh!

Pete

 

[Scotfleiger]

Hi Pete

I have been in contact with the P3D team. They point out that P3Dv3.4 is compiled using .NET framework v4.6 not v4.0.30319 as reported by Event Viewer. 

I have tracked down the LINDA problem to something in the loading of the LINDA libraries that is triggering the crash of P3D. I am debugging further. I found that AS16 for P3D started alone using FSUIPC4.ini [Programs] also triggers the crash. Therefore is not just LINDA that is the cause. I will report progress.

[Pete Dowson]

36 minutes ago, Scotfleiger said:

I have been in contact with the P3D team. They point out that P3Dv3.4 is compiled using .NET framework v4.6 not v4.0.30319 as reported by Event Viewer.

I didn't even know it used .Net Framework. I'm pretty sure FSX and FSX-SE don't.

38 minutes ago, Scotfleiger said:

I found that AS16 for P3D started alone using FSUIPC4.ini [Programs] also triggers the crash.

I use AS16 with P3D, loaded and closed via the FSUIPC4.INI, and it doesn't crash or hang on termination here.

Do you have its DLL loading AFTER FSUIPC4 in the DLL.XML file? If not try moving it. I know there can be problems with FSUIPC getting WX radar details if they are the wrong way around. The latest FSUIPC4 installer makes sure this is right, and I think the AS16 one does too.

Pete

 

[Scotfleiger]

14 minutes ago, Pete Dowson said:

Do you have its DLL loading AFTER FSUIPC4 in the DLL.XML file? If not try moving it. I know there can be problems with FSUIPC getting WX radar details if they are the wrong way around. The latest FSUIPC4 installer makes sure this is right, and I think the AS16 one does too.

The DLL.XML had AS16 loading after FSUIPC4. I reordered them to place FSUIPC4 last but got the AS16 message: Simulator Module Configured - If you sim is already running you must restart it so that the module will load correctly. Afterwards, the AS16 entry was again placed after FSUIPC4. P3D still crashed even if I closed AS16 manually before P3D. 

I have the following DLLs running through DLL.XML: orbx object flow_p3d.dll, Aerosoft As-FlightRecorder.dll, VAInterface.dll, PMDG_HUD_interface.dll, PMDG_Interface.dll and RAASPR.dll.