BuzzLine

Hello there ! I just finished the campaign. *EXCELLENT* work, the concept of "the arctic" was utterly brilliant, I had no idea it was possible to do something like that and I enjoyed the missions there thoroughly. (and before you ask, let me write something on Steam too) I had a minor issue with the mission 13: A suggestion for improvement: the weather does not look that bad (you can even see some blue sky at times), there are strong winds but good visibility, no snow falling. Maybe there could be a way to have "harder" weather ? Anyhow, excellent campaign, great job ! Looking forward to part 2. Keep up the good work.

Patience. It's barely been *4 months* (*) since this was reported. A possible workaround could be to change the dates of all night missions to happen during full moon, maybe. Getting all content creators on board would still be much faster than waiting for a proper fix, apparently. (I don't know if full moon is enough if there is an overcast, though) (*) other thread started on May 2nd :

Happy to help. Tell me if it solves your issue, whenever you get to test it. Cheers ! Buzz

Based on this recent thread (below), it seems the bug is still present in some conditions with ED's code only. I recommend people keep applying my fix. Cheers, Bz

I provide a fix in this thread. Hopefully one day some ED dev will read it... In the meantime, it may fix your issues.

Maybe I can be of help. I have not tested it but, from reading the code, it seems like a small change may solve your issues. The code triggering the promotion seems to be this block below (in file MissionEditor/modules/me_logbook.lua , around line 1266) : if (ProductType.getType() ~= "LOFAC") and (r.name ~= currentPlayer.player.rankName) then currentPlayer.player.rank = r.nativeName currentPlayer.player.rankName = r.name displayPromotionGreeting(r.nativeName) end You want this to never execute. There are several ways. 1. You can simply remove those 5 lines ! (Do make a copy of the file, should you want/need to revert. Or just use the standard "repair" function from the standalone client or Steam, whichever is the way you manage your install) 2. You can replace just the first one by this one here below (I've added "false and", which is easy to revert and prevents the updating code from executing): if false and (ProductType.getType() ~= "LOFAC") and (r.name ~= currentPlayer.player.rankName) then 3. Or comment all the whole block by adding "-- " (that's dash, dash, space) in front of each line. The code is still here should you want to read it but the computer will skip all the lines starting with -- : -- if false and (ProductType.getType() ~= "LOFAC") and (r.name ~= currentPlayer.player.rankName) then -- currentPlayer.player.rank = r.nativeName -- currentPlayer.player.rankName = r.name -- displayPromotionGreeting(r.nativeName) -- end There are other ways but these three are the simplest. I would go for the first option. Brutal but effective. We can call it Code Disappearance Option (CDO). Note : It will work until a future patch updates this file, reverting it to ED's provided version, and then you'll have to put back those modifications... Hope this helps, Buzz. Note : there is another way: changing the threshold at which the rank of captain ends, set it to a value beyond the heat death of the universe. The value lives somewhere in dbCountries but I don't know where that db lives or if it's encrypted or not. Maybe someone knows ? Although I suspect this db might change even more often than that lua file...

Hi Flappie, thanks for the quick answer. I only play with my additional bug protection so I don't know if the original issue is present. And I'm not super keen on flying 1h or 2h campaign mission just to check it. Also, maybe someone has fixed the issue (an undefined id for statics, I guess) but based on ED's track record, there is *no* guarantee of this issue not reappearing. Finally: the current ED code that I touched makes no sense. The "if" I move is either always true or is never reached because it will have crashed three lines earlier. It's like an airbag with a 5s deployment delay ! By the time it triggers, you have already crashed. So I moved that "if" three lines earlier and now it provides additional defense-in-depth should there ever be a target with Nil id. If the ED fix is good and persists in time, good. If not, I have my safety net and I won't stupidly lose two hours of my time and not progress in a campaign. TLDR : my patch is here, feel free to use it. I hope it's never useful (like an airbag) but anyone who wants it feel free to use it. I do recommend having a dev actually look at the code, do a facepalm and chuckle and move that "if" three lines above. Or remove it altogether. It's really silly where it stands:

For some reason the old End Mission thread was locked but the code is *still* brittle and unsatisfactory. So the previously reported "End Mission" bugs may still occurr. If that's the case, here's my additional fix rebased on top of latest open beta 2.7.4.9632 of July 21st (attached). The original discussion can be found here : Hope this helps and may your missions always find an end. Buzz me_logbook.lua

Oooh, we expand to other airframes ? All right then ... Gosh, I thought you americans were cute with your flying plastic toys but seriously ... the Tomcat ?? Puh-lease. Didn't *any* of your airplane designers study aerodynamics ? Is your answer always to simply tack on more or bigger engines ? The *Mirage 2000* is the only plane with both aesthetics and performance. And one that looks like it *actually* belong in the air rather than on a mall's parking lot. There, that should do it.

I think a trigger is broken in mission 2, the one that warns you of incoming missiles. I believe the MISSILE_IN_ZONE call is missing the (mandatory) missile type, it's empty in the original mission. So I've added it (SA2V755, near the bottom of the list) and now it seems the mission has the proper flow. HTH, Buzz Note : I'm not sure this issue is the same as those that are bothering the other two reports of weirdness on that mission, likely but not certain. Feel free to merge threads.

This thread definitely calls for a Dos Gringos moment : (caution, explicit lyrics)

Hey Japo. I've had this issue last time I played R1. And I'm now having it again in the M2000 Red Flag mission. You are not alone. And no, I couldn't find a fix. -sigh- At least when SkyNet rises and the machines become self-aware and wipe out humanity, we'll now they won't be using ED's code.

Here's two tracks : - Huey : NVG_broken_huey_OK.trk Custom mission for my training needs. *NVG working as expected*, this is the control case. - Mi8 : NVG_broken_mi8.trk NVG broken, can't see a thing... That's the mission from Spring Tension with everything remove to fit the 5MB limit of the forum ... (How are you supposed post tracks with this limit ?? What's the trick ? ) Other tracks available but larger than 5MB so I don't know how I can upload them. List of campaign missions where I've seen this so far : - Mi8 Spring Tension, mission 19 (the one where your drop light bombs), also another one but I can't remember which one (convoy maybe ?) - M2k Red Flag, both night missions (the 1st dogfight one and the rocket firing one under mudhen illumination) are very *very* dark => As far as I can tell, half or more of the night missions are currently unplayable ? This needs to be fixed, we fly all-weather ops here !

Thanks for the confirmation. In this case, it could make sense to enable Petro for the ATGM variant of the "Caucasus firing range" instant mission. It's currently off. I'm not sure how I can aim ATGMs and fly at the same time without it?

I think there is a setting in the Mission Editor called "Simplified AI" for the player Mi-24, I'm not sure what it does, but it could enable/disable Petrovich ? It seems that it's not on for some (all ?) of the instant missions, but it defaults to "on" for custom created missions... You may want to change them (already mentioned somewhere in the Mission/Campaign subforum here).

Hello people ! Today saw a great new update on open beta, Cyprus included in the Syria map, and the Hind is a *very* enjoyable beast, so it is a good day to fly. BUT... ...for those who wondered, the buggy lua file was *not* updated, so feel free to continue using my fixed file (linked below), it should keep working and providing additional protection against the bug. Hugs and love from 50ft above Cyprus, Buzz PS: Petrovich says hi to everyone following this bug o/

Yes. If you end a mission with a 'result value' lower than 50, you will regress to the previous mission. At exactly 50 you will replay the mission. At 51 and over, you will advance to the next mission. The 'result value' is defined by the campaign creator and they will "award points" to successful actions, don't confuse it with the "score" value which is only the number of units you killed and is not tied to anything. (It's explained in the DCS User Manual, although it is short, arcane, and easy to miss : "If mission total values are 49 or lower, the player will fall back one stage. If the value is 50, they stay in the same stage. If the mission total value is 51 or above, the player will advance to the next stage. By populating a stage with multiple missions, each with different values, you can create a campaign that flows back and forth according to mission results." ) I believe if you Close the debriefing window instead of clicking on End Mission, then the results will not be taken into account (I may be mistaken, though). I find this mechanism infuriating and I'm happy that most (almost all ?) campaigns I played recently do not use this feature. Instead, they set your score to 50 even if you crash at mission start . Cheers, Buzz

@Roosterfeet Yeah, from what I read, my fix should work fine for your issue. But I'll be damned if I understand why the bug triggers. The debrief.log file looks correct but for some reason, one of the units you killed is not registered in the list of units. This "should not happen" but obviously it does. (Rant: and that's the reason good software engineers use defensive programming... -sigh- ) Anyways, thanks for the log file, it should help the devs actually fix the bug (for real this time). In the meantime, use my fix from above, it should solve (or at least hide) the issue. Cheers ! PS: is Rising Squall any good ? I'm considering buying it...

I concur. I've done my "reproduce the crash" routine several times and so far I have *not* crashed. It's looking good, people. We *may* start to breathe again. Good job Sasha (or whoever fixed the bug)

Same here. But as a software engineer by trade, while I agree that a good debugging story can sometimes read like a novel (BTW, if anyone in software dev is reading this, I recommend Raymond Chen's blog "The Old New Thing" for plenty of epic stories), they are hard to explain in laymen's terms. And the causes are almost invariably the same story : "We found a race condition/memory corruption/uninitialized variable (pick one) and Sasha (statistically, a russian dev is called Oleksandr a.k.a Sasha) found it by bisecting the pile of commits/sheer brainpower/randomly moving debug statements/generously sprinkling mutexes (pick one)". If my guess is right, until "Sasha" finds the cause (maybe causes, there could be several sub-bugs), there is nothing more to say than "We are working on it". And no promises can be made. Once someone is able to reproduce reliably and "Sasha" can send them instrumented builds or intermediate versions, then it's only a matter of time until the bug is identified. Patience ... Waiting is.

Can we roll back to the previous situation ? I prefer to have a crash when (re)starting a mission rather than when finishing one. Right now, OB is not usable for me for any purpose.

And here's a variant of the previous error in initTerrain ... Slightly different callstack. We're making progress ! Or not as my previous post above showed different errors in a different place. So maybe we're regressing, actually... 021-05-27 17:16:54.086 INFO EDCORE: try to write dump information 2021-05-27 17:16:54.088 INFO EDCORE: # -------------- 20210527-171654 -------------- 2021-05-27 17:16:54.089 INFO EDCORE: DCS/2.7.1.6709 (x86_64; Windows NT 10.0.19042) 2021-05-27 17:16:54.090 INFO EDCORE: D:\SteamLibrary\steamapps\common\DCSWorld\bin\dx11backend.dll 2021-05-27 17:16:54.092 INFO EDCORE: # C0000005 ACCESS_VIOLATION at 79D11E68 00:00000000 2021-05-27 17:16:54.094 INFO EDCORE: SymInit: Symbol-SearchPath: '.;D:\SteamLibrary\steamapps\common\DCSWorld;D:\SteamLibrary\steamapps\common\DCSWorld\bin;C:\Windows;C:\Windows\system32;SRV*C:\websymbols*https://msdl.microsoft.com/download/symbols;', symOptions: 528, UserName: 'Romain' 2021-05-27 17:16:54.095 INFO EDCORE: OS-Version: 10.0.19042 () 0x100-0x1 2021-05-27 17:16:54.439 INFO EDCORE: 0x0000000000031E68 (dx11backend): createRenderer + 0x1A0F8 2021-05-27 17:16:54.439 INFO EDCORE: 0x0000000000030B3E (dx11backend): createRenderer + 0x18DCE 2021-05-27 17:16:54.439 INFO EDCORE: 0x000000000003D9F5 (dx11backend): createRenderer + 0x25C85 2021-05-27 17:16:54.439 INFO EDCORE: 0x0000000000008F82 (Terrain): createGlobalLand + 0x272 2021-05-27 17:16:54.439 INFO EDCORE: 0x0000000000009A09 (Terrain): createGlobalLand + 0xCF9 2021-05-27 17:16:54.439 INFO EDCORE: 0x00000000000096FB (Terrain): createGlobalLand + 0x9EB 2021-05-27 17:16:54.439 INFO EDCORE: 0x0000000001C5EEE3 (DCS): AmdPowerXpressRequestHighPerformance + 0xE6FEDF 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000001D40 (lua-terrain): (function-name not available) + 0x0 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000007AC5 (lua): luaD_growstack + 0x845 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000018D94 (lua): luaS_newlstr + 0x5254 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000007E04 (lua): luaD_growstack + 0xB84 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000006F7F (lua): lua_getinfo + 0x133F 2021-05-27 17:16:54.440 INFO EDCORE: 0x000000000000812E (lua): lua_yield + 0x9E 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000002576 (lua): lua_pcall + 0x66 2021-05-27 17:16:54.440 INFO EDCORE: 0x00000000000F2DA9 (edCore): ED_lua_pcall + 0x59 2021-05-27 17:16:54.440 INFO EDCORE: 0x0000000000004005 (lua-dxguiWin): (function-name not available) + 0x0 2021-05-27 17:16:54.440 INFO EDCORE: 0x00000000007D135D (DCS): SW + 0x4E085D 2021-05-27 17:16:54.441 INFO EDCORE: 0x0000000001C5C86D (DCS): AmdPowerXpressRequestHighPerformance + 0xE6D869 2021-05-27 17:16:54.441 INFO EDCORE: 0x0000000000A07E9E (DCS): SW + 0x71739E 2021-05-27 17:16:54.441 INFO EDCORE: 0x0000000000017034 (KERNEL32): BaseThreadInitThunk + 0x14 2021-05-27 17:16:54.441 INFO EDCORE: 0x0000000000052651 (ntdll): RtlUserThreadStart + 0x21 2021-05-27 17:16:54.734 INFO EDCORE: Minidump created. 2021-05-27 17:16:54.734 INFO Lua::Config: stack traceback: [C]: in function 'Init' [string "./MissionEditor/modules/me_map_window.lua"]:817: in function 'initTerrain' [string "./MissionEditor/modules/me_mission.lua"]:1637: in function 'load' [string "./MissionEditor/modules/me_quickstart.lua"]:274: in function 'func' [string "./Scripts/UI/ProgressBarDialog.lua"]:154: in function 'updater' [string "./Scripts/UpdateManager.lua"]:64: in function <[string "./Scripts/UpdateManager.lua"]:40> 2021-05-27 17:16:54.734 INFO Lua::Config: stack traceback: 2021-05-27 17:16:54.734 WARNING LOG: 1 duplicate message(s) skipped. dcs.log

[UPDATE : the callstack in this error is different and comes from a different place than before the patch 2.7.1.6709] I got the ESC crash (or shift-R crash, I have both). Please find attached two logs, one for a session that works and one for a session that crashes on esc or shift-R. The working one also has the "regenerate fxo" shaders step as I had purged them. Then I try to load one instant mission and either shift-R or escape then quit then do another one. I don't know if purging fxo each time is a solution, I'll try. When I diff the two files, the first significant difference between the two files seems to be a VFS error, and then it dumps. 2021-05-27 16:43:34.025 DEBUG LuaGUI: ----onSimulationStop--- false nil nil 2021-05-27 16:43:34.025 DEBUG LuaGUI: ----DCS.isMultiplayer()--- false 2021-05-27 16:43:34.054 INFO Config: netview stopped 2021-05-27 16:43:34.059 ERROR VFS: VFS_open_write: CreateFile(): Unknown error <<<<<<<< This does not happen in the "good" log 2021-05-27 16:43:34.059 ERROR DCS: can't open file 2021-05-27 16:43:34.072 INFO TERRAIN: lSystem::CleanScenes() 2021-05-27 16:43:34.075 INFO VISUALIZER: StopSimulation 2021-05-27 16:43:34.076 INFO VISUALIZER: Stopped collection of statistic. 2021-05-27 16:43:34.089 DEBUG LuaGUI: --- onChatMessage--- Mission is over. nil 2021-05-27 16:43:34.090 INFO EDCORE: try to write dump information Note : there's also a debug output that's present in the good log and missing in the bad, just a few lines above the VFS error : 2021-05-27 16:39:05.860 DEBUG LuaGUI: ----onSimulationPause--- HTH, Buzz dcs_CRASH.log dcs_OK.log.old

Before someone asks: there was no change from ED to the file in 2.7.1.6709 from today's update. So my fix (above) will continue to work on that version and (maybe) bring some additional robustness. At least it should not harm anything. Enjoy.

@BIGNEWY : if you want to send over a fully instrumented build so I can crash it and send you back detailed stack traces and log trails, I'll be happy to help.