Autonomous CAP and GCI AI fighter script

[Scooternutz]

I can verify that the GCI script is working without error again after today's patch.

 

I concur.

 

Update: Problem seems to return 20 to 30 minutes into the mission.

Edited December 30, 2016 by Scooternutz

[Scooternutz]

Since last update I have been getting a crash about 20 to 30 minutes into the mission. Please see crash log, thanks!

 

 

 

# -------------- 20161230-195638 --------------

�

# C0000005 ACCESS_VIOLATION at 5E2BD787 00:00000000

00000000 00000000 0000:00000000

5E2BD787 00AFEA60 0000:00000000 ?isGroup@wControl@@QEBA_NXZ()+7

6B72D3D4 00AFEAA0 0000:00000000 ?regLuaEnv@Scripting@@YAXPEAUlua_State@@_N@Z()+8B4

6E857254 00AFEAE0 0000:00000000 luaD_growstack()+7A4

6E866FB5 00AFEC60 0000:00000000 luaS_newlstr()+49C5

6E857541 00AFEC90 0000:00000000 luaD_growstack()+A91

6E85684F 00AFEE10 0000:00000000 lua_getinfo()+11AF

6E85785E 00AFEE50 0000:00000000 lua_yield()+9E

6E86C8E3 00AFEEA0 0000:00000000 luaL_newstate()+20C3

6E857254 00AFEEE0 0000:00000000 luaD_growstack()+7A4

6E866FB5 00AFF060 0000:00000000 luaS_newlstr()+49C5

6E857541 00AFF090 0000:00000000 luaD_growstack()+A91

6E85684F 00AFF210 0000:00000000 lua_getinfo()+11AF

6E85785E 00AFF250 0000:00000000 lua_yield()+9E

6E852440 00AFF2A0 0000:00000000 lua_pcall()+60

6B73A8BD 00AFF2D0 0000:00000000 ?regLuaStaticObject@Scripting@@YAXPEAUlua_State@@@Z()+37D

704D642F 00AFF350 0000:00000000

704D6975 00AFF3A0 0000:00000000

53326CF8 00AFF410 0000:00000000

5332E4C1 00AFF470 0000:00000000

5331C354 00AFF4A0 0000:00000000

5331C2A4 00AFF4D0 0000:00000000

52FF743F 00AFFB80 0000:00000000

53432ABD 00AFFBC0 0000:00000000

80808364 00AFFBF0 0000:00000000 BaseThreadInitThunk()+14

811970D1 00AFFC40 0000:00000000 RtlUserThreadStart()+21

 

 

[Stonehouse]

Not really script errors per se. Seems similar to the errors recently seen in the autonomous ground force script by gromit190 https://forums.eagle.ru/showpost.php?p=2997393&postcount=23

 

I can't decipher what you are showing - which is basically a dump listing - all I can say really is that I think DCS is trying to write to a location in memory that is locked/unaddressable and so fails.

 

I would suggest making a copy of your mission and take out everything but the GCICAP set up so it is just the bare bones and then test if you still get a crash at the same point. If you do it is probably worth opening a post in the bugs thread as it may well be a DCS bug or something has changed in the last few patches to render GCICAP unworkable in it's current version. Putting in a bug thread with the logs and mission attached may get other more knowledgeable people involved.

 

If you don't crash then it implies that some of the other objects or triggers or scripts etc etc are causing the problem so you could then add them back a few at a time noting what you add each time and after each addition run the mission. If it crashes then something in the last added set is the culprit. If you wanted to then you could remove the last set of things and add them one at a time noting each until you get the crash bug. This would imply that possibly the last thing you added was the cause or perhaps the last thing added creates a fatal combination. In any case this would be more info for your bug report and help the ED guys locate and fix the issue. I guess you could also PM lukrop to see if he is able to assist you.

[Pikey]

I saw those in the last version, but not the current one yet. However, as SH says, we've seen these throughout times in 1.5, its the tasking of units crapping out with a memory error and its a DCS limitation currently. Try to task less units at the same time, it's usually a sign that you are reaching the glass ceiling in DCS and it's something that makes me sad every time I see it. Both CTLD and GCICAP and recently the ground forces movement can aggravate it.

 

In my opinion, nailing the performance of moving and ordering units in MP would be the single biggest improvement to DCS for MP we could get, Blue Flag proves it as it straddles the glass ceiling of scaling and they worked it for years now.

[S D]

Can confirm, we are getting the same crashes. Even on my reduced load version. :(

[Stonehouse]

OK if the crash is consistent and repeatable with a simple pure GCICAP only set up probably start a possible bug thread in https://forums.eagle.ru/forumdisplay.php?f=487 and attach your logs and the mission and the details. I did see that there was a recent hotfix to mist so just be sure that you are using the latest one before you post. Pikey/ gromit190 if you can do the same thing with the autonomous ground script and/or CTLD in his thread that would probably help ED see that it isn't just a GCICAP issue.

[Chicki]

Stonehouse, you mind saying exactly WHERE you saw this hotfix to MIST? Appreciate it.

[Stonehouse]

Just the usual mist github site https://github.com/mrSkortch/MissionScriptingTools/releases/tag/4.3.74

 

Though looking at it properly now it was sept 2016 so you should have seen it already but it happens sometimes that people don't keep up to date. Know it's caught me once or twice.

[Chicki]

Naw pal. I'm just not that smart. Will keep an eye out now on though. Like to try it on WDrago's mission. It seems to want to crash after about 30 min or so. Worked fine until these last couple updates.

 

https://www.digitalcombatsimulator.com/en/files/1726812/

[S D]

Can confirm crash happens with latest MIST version, my template was built from scratch using the latest for the Spitfire release.

[Stonehouse]

OK considering it used to work without crashing and that the GCICAP script has not been updated the logical implication is that DCS changes are connected to the problem.

 

So as I said people should create a thread over in the release bug area to ask if it can be investigated to confirm it is a bug and if so rectify it.

 

However please be sure to have no mods loaded when generating the logs of the crash and the mission itself as having any mods on will obscure the matter. ie raise questions as to whether it is the mod or DCS possibly being the root cause. Obviously attach the mission and logs to your post in the bug sub forum. Think Big Newy has a link to guidelines for bug reporting in his sig.

[Pikey]

I picked up an older mission with Mist 4.3.73 and ran it for 2 hrs at x20 without it crashing. Is this significant or was i just lucky? I dont have issue with it on NTTR either.

[Stonehouse]

Not sure. Can you make a copy of it, edit it and make your only change be updating to Mist 4.3.74 and then try to run the edited mission and see what happens? It may be Mist related and not DCS after all.

[71st_AH Rob]

Not sure. Can you make a copy of it, edit it and make your only change be updating to Mist 4.3.74 and then try to run the edited mission and see what happens? It may be Mist related and not DCS after all.

 

I have been troubleshooting this problem for a while now, I was running 4.3.73 when the problem appeared, I have been disabling scripts one at a time and updated MIST thinking it was the problem, the only thing in common is running GCICAP. There is no rhyme or reason to when the mission seems to crash but it always does when the script is enabled, sometimes on start-up, sometimes after several hours

[Pikey]

Out of interest, how many caps are you spawning? How many aircraft total and what types? I can troubleshoot someones mission if they want?

I have been troubleshooting this problem for a while now, I was running 4.3.73 when the problem appeared, I have been disabling scripts one at a time and updated MIST thinking it was the problem, the only thing in common is running GCICAP. There is no rhyme or reason to when the mission seems to crash but it always does when the script is enabled, sometimes on start-up, sometimes after several hours

[71st_AH Rob]

Out of interest, how many caps are you spawning? How many aircraft total and what types? I can troubleshoot someones mission if they want?

 

I am using 3 CAP flights on each side with borders on. Blue is 4 a/c per flight and Red is 2 or 4 a/c per flight.

 

Ran it once with MIST 4.3.74 and it ran for hours with no problems, GCICAP was the only script running, then ran it again, crashed on start up.

[Pikey]

been trying for a while but I did get a CTD on .74. Not had one on 73 yet.

[71st_AH Rob]

been trying for a while but I did get a CTD on .74. Not had one on 73 yet.

 

Ill try going back to .73 again when i get back in town but I don't expect anything different. The crash seems to be fairly random.

[S D]

Tried reverting back to .73 with the same results.

 

Here is my mission if you would like to check it your end Pikey. Currently its got .74 in, but you can swap that out.

 

Crash seems pretty consistant, around the 15/20 minute mark.

TemplateReduced.miz

[Pikey]

Hi, reproduced your findings on live, same memory error as we see in live, at about the point you said. i don't think this has anything to do with the scripts or version of MIST per se. I believe after seeing it a lot, that its a tasking issue from the DCS scripting engine. There's a possibility that a lot of your units are getting RTB at the same time....its about that point. I did the same with ground units and see the same error.

 

My recommendations are to lower the plane count. 18 on the map and my PC didnt update well even at 10x. What you are hitting is the DCS glass ceiling. To be fair, that's over a full squadron in the air at one time, so it's pretty WW3 if you catch my drift and thus you wont sacrifice any realism knocking the numbers down. HTH.

Tried reverting back to .73 with the same results.

 

Here is my mission if you would like to check it your end Pikey. Currently its got .74 in, but you can swap that out.

 

Crash seems pretty consistant, around the 15/20 minute mark.

[Pikey]

I'm willing to bet a pint that if you did the same on NTTR 2.04 it wouldn't crash. :)

 

I don't know how to meaningfully provide this as feedback to the moderators/testing team.

[Scooternutz]

I have attached my mission. I can't figure out why it crashes from 15 to 30 minutes in. I reverted to MIST 73 from 74. I only have two CAPS and GCIs.

 

I am at a loss.

16th_ACCW_CAS_GCICAP_Delta6.miz

[Stonehouse]

It's almost at the point of putting things on hold until the versions merge or at least until Normandy comes out. Pretty much, beyond reporting the problem and waiting for the dev team to fix the issue, I don't think there is much to be done as I don't believe the root cause is in the scripts.

[Scooternutz]

It's almost at the point of putting things on hold until the versions merge or at least until Normandy comes out. Pretty much, beyond reporting the problem and waiting for the dev team to fix the issue, I don't think there is much to be done as I don't believe the root cause is in the scripts.

 

Agreed. :(

[Pikey]

I spoke with BIGNEWY about reporting this. There is a difficulty in that we don't get traction for two reasons. Firstly its the moving target scenario where ED are concentrated on getting to 2.5. Second its that we supply a reproduction that only involves an external script with behaviour ED doesn't anticipate as a core function of it's product.

 

I've written in another post, my actions personally is that I'm not pursuing a fix on this version, given the timing and time of year it happened and the larger picture. But I am watching.

 

In the meantime, reduce the CAP's. it works.