resolved 2.7.0.4625 Open Beta crash at second mission load

[Flia]

At least WIn 7 and Win 8 Compatibility mode still helps. I just tested DCS with compability off and i crashed immidietly whwn i pressed Shift+R. 

Edited May 16, 2021 by Flia

[Flappie]

21 minutes ago, petsild said:

Hi Flappie,
can you please provide your log file dcs.log thank you.  Pet

I can do it, yes. What mission do you want me to start (aircraft, mission)? Do you have full log enabled or disabled? If you want full log, copy this file to your "Saved Games\DCS...\Config\" folder.

autoexec.cfg

Edited May 16, 2021 by Flappie

[wernst]

6 minutes ago, Flia said:

At least WIn 7 and Win 8 Compatibility mode still helps. I just tested DCS with compability off and i crashed immidietly whwn i pressed Shift+R. 

 

just to understand: Does W7 or W8 comp modes give you zero crashes or do they help getting less crashes?

How long have you tested the W7 and W8 comp modes?

[the_jovian]

1 hour ago, Ironhand said:

 

Assuming that fix is what was pushed to my computer yesterday (?), I experienced a crash immediately after installing it.

 

Yeah, although I originally thought that the WIndows patch helped, I don't think it actually did anything at all related to this.

 

1 hour ago, hornblower793 said:

Aren't we all, but short of getting a dev build or full debug build on a problem machine I don't what else we can do - takes me back to my time leading development teams building mission critical systems for police and airport clients where some of the 'features' took months to track down
 

 

Here are some theories that I have on what might be happening based on previous hard to pin down crashes. I think the key issue is that only some users are affected. Therefore it has to be something in these specific user's machines that is different than the dev teams. So what could it be?

• Libraries: if the Windows 8 compatibility actually works, then this might be the case. I've seen strange crashes when clients have had old/different versions of .NET environment installed. I don't think this could be the case here, but maybe another library? See also https://superuser.com/questions/133746/how-does-the-compatibility-mode-in-windows-work
• Localisation: I've seen strange crashes when there is a conflict in, e.g., what decimal separator is used. For example, one library pulling it from Windows region settings and another assuming a period. Then you could, e.g., get a conflict between comma and a period, leading to string-number conversion errors. This could then lead to a crash if you try using the result in dynamic memory allocation. I just remembered that I'm using a comma as the separator so I'll try switching it to a period for giggles. [edit]: tried this, no luck
• Environment, like installation paths or usernames. Sometimes I've seen bugs where a fixed array just overflows if the path name is unexpectedly long. But my username isn't long at all. Maybe something different?
• Files gathering somewhere in the system causing some type of overflow. That's what we're trying to fix with deleting the temporary archives. I haven't really had much success with this though. I wonder if we've identified all the temporary folders where DCS is writing stuff?
• Drivers: but this crash seems to be happening with all kinds of configurations. Maybe we all share some kind of a (virtual?) peripheral that's causing this? For example I'm using vJoy with Joystick Gremlin, could something like that be causing it?
• Settings: but all kinds of driver and DCS settings have been suggested with no luck.
• Modules: I have a feeling that the F-14 is particularly crashy, but could it be that there's some module which if installed is causing this?

One possibility to keep in mind is that when you compile code with a non-release configuration, there's might be some extra stack space available. Also variables might get initialized to zero which won't happen in production. I've had this happen a couple of times, see, eg., https://stackoverflow.com/questions/186237/program-only-crashes-as-release-build-how-to-debug .I'm sure there's a bunch of other differences as well. But I think the closed beta testers like @Flappie will have a release build, so this seems pretty unlikely.

 

I'm not a professional software developer by any means. Can you think of any other avenues of investigating this?

Edited May 16, 2021 by the_jovian

[wowbagger]

There's no way to move the TEMP folder that DCS uses is there? That is, have it use a different location.

[Flia]

16 minutes ago, wernst said:

just to understand: Does W7 or W8 comp modes give you zero crashes or do they help getting less crashes?

How long have you tested the W7 and W8 comp modes?

So far 0 crashes with Win7 or Win 8 Compatibility modes. I loaded mission, restaretd mission many times. Lets say i have at least 20 or 30 tests with no crash at all. Even with not deleted fxo files.

Edited May 16, 2021 by Flia

[Tanuki44]

On 5/10/2021 at 4:01 PM, Tanuki44 said:

I checked 'Compatibility Windows 8' in DCS.EXE properties.... 

 

For the moment, no crash ....

I started on 5/10 and no crash since, unless I disable compatibility, but this is only a solution in the meantime.

[petsild]

59 minutes ago, Flappie said:

I can do it, yes. What mission do you want me to start (aircraft, mission)? Do you have full log enabled or disabled? If you want full log, copy this file to your "Saved Games\DCS...\Config\" folder.

autoexec.cfg 44 B · 1 download

 

Just dcs.log selection of mission anything, thank you.

[wernst]

2 hours ago, Flia said:

So far 0 crashes with Win7 or Win 8 Compatibility modes. I loaded mission, restaretd mission many times. Lets say i have at least 20 or 30 tests with no crash at all. Even with not deleted fxo files.

 

It's just unbelievable. Let's say that's the solution. Then the question must be:

What is different from W10 to W7 or W8 that prevents the crashes in DCS?

And what is different from 2.5.6 stable to 2.5.7.xxxx which seems to doesn't "like" the W10 but more the W7 or W8? (just to use simple words) 

Edited May 16, 2021 by wernst

[petsild]

I returned all my mods  with the enabled Compatibility modes W7, the third day without crashes.

[wernst]

4 minutes ago, petsild said:

I returned all my mods  with the enabled Compatibility modes W7, the third day without crashes.

 

Did you have the "2nd load crash" before you switched to W7 comp mode? 

[Flappie]

31 minutes ago, petsild said:

Just dcs.log selection of mission anything, thank you.

Here you go.

dcs.log

[petsild]

 

thanks

[petsild]

51 minutes ago, wernst said:

 

Did you have the "2nd load crash" before you switched to W7 comp mode? 

In W8 / W7 mode, no crash ever.

[Wali763]

2 hours ago, wernst said:

just to understand: Does W7 or W8 comp modes give you zero crashes or do they help getting less crashes?

How long have you tested the W7 and W8 comp modes?

I have for sure 150+  reloads in W8compmode without a single crash now. Disable it and the crashes are back...

 

I dont care about the longer loadtimes right now.

[Daytona675]

W8compmode has been working for me too. Have been practicing and crashing a lot doing carrier landings. 

[the_jovian]

I'm going to try the W8 compatibility mode too. Initial tests look promising...

 

When DCS crashes, it dumps out the loaded modules in the following way:

 

2021-05-15 22:04:29.929 INFO    DCS: enumerating loaded modules
2021-05-15 22:04:29.929 INFO    DCS: C:\Program Files\Eagle Dynamics\DCS World OpenBeta\bin\DCS.exe, 21838176 bytes, created 14:4:2021-15.47.58, written 14:4:2021-15.47.58, accessed 15:5:2021-22.4.30
2021-05-15 22:04:29.929 INFO    DCS: C:\Windows\SYSTEM32\ntdll.dll, 2024728 bytes, created 29:4:2021-15.34.41, written 29:4:2021-15.34.41, accessed 15:5:2021-22.4.30

 

Is there any way to get the same output from DCS when the game doesn't crash?

[petsild]

1 hour ago, Flappie said:

Here you go.

dcs.log 59.39 kB · 6 downloads

My log is almost 3x longer, no DEBUG line for you?

 

I have a set of bugs and friends have similar:

 

2021-05-15 19:37:37.735 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/keyboard/Keyboard.lua'

2021-05-15 19:37:37.736 DEBUG LuaGUI: Input: File[./Config/Input/UiLayer/keyboard/default.lua] opened successfully!

2021-05-15 19:37:37.737 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/keyboard/Keyboard.diff.lua'

2021-05-15 19:37:37.737 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/joystick/MJ16.lua'

2021-05-15 19:37:37.737 DEBUG LuaGUI: Input: File[./Config/Input/UiLayer/joystick/default.lua] opened successfully!

2021-05-15 19:37:37.738 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/joystick/MJ16.diff.lua'

2021-05-15 19:37:37.739 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/joystick/ VKB-Sim Gladiator NXT R .lua'

2021-05-15 19:37:37.739 DEBUG LuaGUI: Input: File[./Config/Input/UiLayer/joystick/default.lua] opened successfully!

2021-05-15 19:37:37.740 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/joystick/ VKB-Sim Gladiator NXT R .diff.lua'

2021-05-15 19:37:37.740 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/trackir/TrackIR.lua'

2021-05-15 19:37:37.741 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/trackir/default.lua'

2021-05-15 19:37:37.741 DEBUG LuaGUI: Input: can't open './Config/Input/UiLayer/trackir/TrackIR.lua'

2021-05-15 19:37:37.741 DEBUG LuaGUI: Input: can't open

It continues...

dcs.log

[Flia]

1 hour ago, wernst said:

It's just unbelievable. Let's say that's the solution. Then the question must be:

What is different from W10 to W7 or W8 that prevents the crashes in DCS?

And what is different from 2.5.7.xxxx to 2.5.6. stable which doesn't "like" the W10 but more the W7 or W8? (just to use simple words) 

I agree. This problem and his solution is really unbelievable

[84-Simba]

Still happens from time to time.

One observation though :

 

Let's say I fly a mission called Hornet Trap Case 1 and quit.

 

Sometimes, back in the ME, on the bottom left I see Hornet Trap Case 1.miz ... and sometimes it says something like missiontemp.miz or something like that.

 

Since you guys talked about the temp folder maybe it could be useful.

[Ironhand]

@Flappie

 

So...looking at what we've gotten so far with this experiment, it seems the fault is occurring either prior to or during the exit from the first mission. The necessary BIN file isn't written properly for some reason. The sim uses this BIN to start the 2nd mission and that's when things go wrong. What we don't know--and probably will never discover--from these files is why it's being truncated/corrupted. But it occurs at the end of the mission preceding the crash. The fault is already in place as you start the 2nd mission.

 

Based on my own experiences and, as I review this thread, for many of us there is a disturbing periodicity occurring. The cycle in my case is roughly 24 hours. I had observed it earlier and, now, with a separate Windows logon for the sim while we've been experimenting it's been very obvious. I'll have the second mission crash immediately after starting the sim, followed the next day by not crashing unless and until I restart the sim after flying for a period of time without an issue. Rinse and repeat.

 

I think this periodicity is why we are seeing so many: "Eureka! I did "X" and I think it's solved. No crashes in the last 20 missions!" Followed in the next day or two by: "It's crashing like crazy again. So I tried "Y" and now it's not crashing. I think it's fixed." Followed by...

 

Don't know what this means but I know it's nothing we're going to solve. And I'm not sure there's any more data we can generate. Perhaps the devs have some suggestions for us to try. If not, then I'm just going to enjoy the sim as best I can and hope that things sort themselves out while I'm still alive. But if there's anything I can do to help, just let me know.

Edited May 16, 2021 by Ironhand

[wernst]

1 hour ago, Ironhand said:

. . . I'm just going to enjoy the sim as best I can and hope that things sort themselves out while I'm still alive . . .

 

No way, I don't do that. I enjoy DCS 2.5.6. stable (in my case 100 % crash free so far)

I enjoy the good things of DCS, even without spectacular clouds.

When I read what recipes are suggested here to avoid crashes, I get even more gray hair . . . while I'm still alive. 

(btw, I have 2.7. on a second drive for test purpose and follow up) 

Edited May 16, 2021 by wernst

[Screenchecker]

Can confirm that I experience the issue in role selection on multiplayer servers (in VR) and that Windows 7 compatibility mode seems to fix it so far. Can't say if it has a performance hit since I have no data to compare, just in-game frame time overlays, but since the numbers generally fluctuate a lot, it's hard to compare. At least the hit is not super obvious. Also had a look into my logs, nothing out of the ordinary in there after the crashes (sorry, I know that is no great help to reproduce the issue). It also never occurred instantly in my case. Usually within the first 30 minutes of a session.

Even though you guys have to take a lot of flak from angry players I gotta tell you that this game brings me a lot of joy. In VR it's a dream come true for flight enthusiasts. 

[petsild]

30 minutes ago, wernst said:

 

No way, I don't do that. I enjoy DCS 2.5.6. stable (in my case 100 % crash free so far)

I enjoy the good things of DCS, even without spectacular clouds.

When I read what recipes are suggested here to avoid crashes, I get even more gray hair . . . while I'm still alive. 

(btw, I have 2.7. on a second drive for test purpose and follow up) 

 

 

In the last 2.7 there are so many improvements and fixes, staying at 2.5.6 is unnecessary.

The compatibility mode will not limit you in any way, it only adds 100% stability.

I worked on the W7 system without a problem for several years until the last update.

[Flappie]

2 hours ago, Ironhand said:

So...looking at what we've gotten so far with this experiment, it seems the fault is occurring either prior to or during the exit from the first mission. The necessary BIN file isn't written properly for some reason. The sim uses this BIN to start the 2nd mission and that's when things go wrong. What we don't know--and probably will never discover--from these files is why it's being truncated/corrupted. But it occurs at the end of the mission preceding the crash. The fault is already in place as you start the 2nd mission.

Are you sure about this? Because it seems the 3 .bin files are only re-used when hitting Shift+R.

When doing "Exit" and then "Fly again", new .bin files are made. Doesn't DCS also crash when doing "Exit" then "Fly again" ?

 

I'm not even sure the issue lies within the .bin files.