Unable to recover, return to KSC, revert, switch vessels, SSUICamera's fault?

[Kerbas_ad_astra]

(KSP 0.25, Windows x32, Steam)

I've recently added and updated a few mods on my install and took a slightly-modified (plus RCS, plus Kerbal Engineer, plus docking port, minus landing legs) Kerbal X up for a spin to see how it worked (in particular, I was trying to use the Trajectories mod to get a precise landing back at KSC). Everything seemed to be going well until I tried to recover the vessel after landing -- no scene-switching buttons worked (recover and return to space center buttons in the altimeter box, space center and revert buttons in the pause menu, "switch-to" button when clicking on other vessels in the map view). There are some early exceptions (e.g. NullReferenceExceptions caused by "Clouds.Clouds.Update()", and what seems like a missing image file in HaystackContinued, which I've reported in that thread), but the part of output_log.txt where I try to switch scenes is as follows:

[HighLogic]: =========================== Scene Change : From FLIGHT to SPACECENTER (Async) =====================

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

NullReferenceException: Object reference not set to an instance of an object

at PatchedConicRenderer.OnSceneSwitch (GameScenes scn) [0x00000] in <filename unknown>:0

at EventData`1[GameScenes].Fire (GameScenes data) [0x00000] in <filename unknown>:0

at HighLogic.LoadScene (GameScenes scene) [0x00000] in <filename unknown>:0

at VesselRetrieval.onVesselRecoveryRequested (.Vessel v) [0x00000] in <filename unknown>:0

at EventData`1[Vessel].Fire (.Vessel data) [0x00000] in <filename unknown>:0

at AltimeterSliderButtons.recoverVessel () [0x00000] in <filename unknown>:0

at ScreenSafeUIButton.OnMouseUp (Boolean rightClick) [0x00000] in <filename unknown>:0

at ScreenSafeUIButton.OnMouseUp () [0x00000] in <filename unknown>:0

at SSUICamera.DoMouseEvents () [0x00000] in <filename unknown>:0

at SSUICamera.Update () [0x00000] in <filename unknown>:0

(Filename: Line: -1)

There are several sections like this in the log, for each time that I try to click the "recover" or some other scene-switching button. The whole log is here: http://www./view/6w0qsrck10dcl17/output_log.txt

I'm not the first person to have grief with SSUICamera.Update(), as seen in this thread and this Pastebin, but I've never encountered this error before -- most of the time I get CTDs.

And now, the confounding factors. I realize that these complicate support, and I will attempt to chase these out as best I can over the next few days, but it will take some time and I want to get as much information as possible about this bug written and posted before I forget it.

• This happened in a fairly well-used save, with a probe on its way to solar orbit (which seemed to have some very strange interactions with Enhanced Navball, which I caused me remove it -- funnily enough, I started this latest session to see if that problem still occurred without ENB, but I got sidetracked testing Trajectories) and several SCANSat and Karbonite mapping satellites around Kerbin, the Mun, and Minmus.
  
• Even with Active Texture Management, I was actually running basically at 3.5 gigs of memory when I force-quit KSP. That's never happened to me before (though I once broke through it completely and crashed during loading), so it might be a factor.
  
• I use quite a few mods, and had updated or added a dozen of them between a few days ago (when I last played) and now (though I am quite sure that all of my mods are up to date -- between KSP-AVC and my subscriptions in KerbalStuff, I think I've kept on top of that pretty well). The ones I changed or added are:
  
  • Enneract's RealRoster 0.25 rev1
    
  • CrossFeedEnabler 3.1, with the FuelWings cfg file (LF+O version, not in play here as the Kerbal X has no wings)
    
  • Trajectories 0.4.1 (I had previously installed this version, but removed when it seemed to be causing crashes [and doing this did seem to resolve the issue]; I added it back again to see if maybe it had been badly interacting with other stuff I've since removed, and it seemed to work like a charm just now)
    
  • FinalFrontier 0.5.9-177 (I had previously installed an earlier version -- 0.5.6b, maybe -- and removed it when it seemed to be causing crashes and writing excessively to the log and to its files, but I put in the latest version because [according to the changelong] it does less of that now.)
    
  • WernherChecker 0.3.1.0
    
  • Aligned/Formatted Currency Indicator 1.2
    
  • Rover Wheel Sounds 1.2
    
  • Rover Sound (and light!) Effects 4.2
    
  • Navball Docking Alignment Indicator 5
    
  • Community Resource Pack 0.2.3 (upgraded from 0.2.2, which worked just fine)
    
  • TextureReplacer 1.8.1 (upgraded from 1.7.4, which worked fine)
    
  • UniversalStorage 0.9.3.25 and US-KAS 0.9.0.14 (upgraded from 0.9.0.whatever the previous version was, which had worked fine)
    

  [*]Other mods that I have installed previously:

  • Kerbal Alarm Clock 3.0.4.0
    
  • KSP Alternate Resource Panel 2.6.1.0
    
  • Transfer Window Planner 1.1.2.0
    
  • Time Control 13.2
    
  • Umbra Space Industries core 0.2.3, MKS/OKS 0.21.2, Exploration Pack 0.2.4, Karbonite 0.4.2, Karbonite Plus 0.2.2
    
  • ScanSat 8
    
  • Resource Overview 1.0.1.1
    
  • Ship Manifest 3.3.2b
    
  • Connected Living Spaces 1.0.11.0
    
  • Haystack Continued 0.3.2.0
    
  • ModuleRCSFX 3.1
    
  • PlanetShine 0.2.2
    
  • CITUtil and KERT (added by USI Exploration Pack)
    
  • EnvironmentalVisualEnhacements 7-4 with Astronomer's Visual Pack V3 "Edge of Oblivion"
    
  • Procedural Fairings 3.10
    
  • RLA Stockalike 12
    
  • StripSymmetry (1.4? I downloaded that one fairly recently...)
    
  • Hooligan Labs Airships 3.0
    
  • MechJeb2 (2.4.0? I downloaded that one fairly recently...)
    
  • Aviation Lights (3.6? I downloaded that one fairly recently...)
    
  • Procedural Wings 0.9.1
    
  • Flag Decals 2.1
    
  • Kerbal Joint Reinforcement 2.4.4
    
  • SelectRoot Jul18
    
  • Vessel Viewer (0.6? I downloaded that one fairly recently...)
    
  • ALCOR 0.9
    
  • RasterPropMonitor 0.18.3
    
  • FAR 0.14.3.2
    
  • TweakableEverything 1.5
    
  • RCS Build Aid 0.5.2
    
  • Active Texture Management (3-8? I downloaded that one fairly recently...it would be nice if more mods included the version number in their directories)
    
  • TAC Fuel Balancer 2.4.1.5
    
  • Radial Engine Mounts 0.30
    
  • Kronal Vessel Viewer 0.0.4_0.25
    
  • Hullcam VDS 0.32
    
  • Vanguard Technologies 0.7.2 and EVA Parachutes 1.2
    
  • Chatterer 0.7.1
    
  • Editor Extensions 1.4
    
  • Firespitter 6.3.5 (KSP-AVC whines at me about it being compiled for 0.24.2, and I know there's an updated DLL floating around -- I'll see about updating it during all this debugging)
    
  • Custom Asteroids 1.1.0
    
  • Docking Port Alignment Indicator 4.0
    
  • ORSX 0.1.2
    
  • Kerbal Attachment System (0.4.9? I downloaded that one fairly recently...)
    
  • Infernal Robotics 0.19.2
    
  • TweakScale (1.44? I downloaded that one fairly recently...)
    
  • PartCatalog 3.0 RC8
    
  • Blizzy's Toolbar 1.7.7
    
  • SurfaceLights...the version hasn't changed in a while, so who cares what the number is?
    
  • ModuleManager 2.5.1
    

Ugh. I feel like I spent more time writing this post than I did playing KSP today!

My next step is going to be to dump my mods into the time-out folder see if I can recreate the issue with no mods and a new save, and then start moving mods in from time-out until I can re-create it again (or until all of the mods are back and the issue doesn't return -- blasted Heisenbugs), but it may be some time before I am done with this. If anyone has insight as to what of the dozen mods in the upper list use the SSUICamera class, I'd be much obliged -- it's probably an issue with one of them. Thanks.

EDIT: So, I just yanked out all of the mods except for the top dozen (and ModuleManager), and executed the same attempted landing at KSC. Everything worked hunky-dory (in fact, I landed at KSC instead of falling short because the stock drag model isn't nearly as sensitive as FAR) except for RealRoster -- I knew that the button wouldn't appear without the toolbar mod, but my crew roster didn't seem to be randomized at all, and a search of the output_log.txt (available here: http://www./view/du7l72cxq4x8khq/output_log_2.txt) showed that it was generating lots of NullReferenceExceptions (and I think it was the only thing generating NREs) like this one:

NullReferenceException: Object reference not set to an instance of an object

at RealRoster.RealRosterGUI.AddToolbarButton () [0x00000] in <filename unknown>:0

at RealRoster.RealRosterGUI.init () [0x00000] in <filename unknown>:0

at RealRoster.RealRoster.Start () [0x00000] in <filename unknown>:0

(Filename: Line: -1)

In contrast, in my heavily-modded save, RealRoster made only a couple of NREs, but it was the first one in the entire log. For my next test, I'm going to put blizzy's toolbar mod back, to see what happens -- maybe RealRoster keeps trying to make a button even when the required toolbar mod isn't present.

Edited November 4, 2014 by Kerbas_ad_astra
Results of first test...

[Kerbas_ad_astra]

In these error stack traces, did the error start on top or on the bottom? I ask because my NREs in the first session (where I found myself unable to recover, return to KSC, revert, or switch vessels) had patchedConicRenderer at the top, as did the log in the pastebin I linked, but different from the other thread on this forum. Unfortunately, a search of the forums and the KSP bug tracker doesn't reveal any other mentions of that pastebin, so I don't know who had that bug or how they got it.

If the error started in patchedConicRenderer, could that mean that it was caused by Trajectories or by some stock bug?

I just did another test on a clean save with only blizzy's toolbar and RealRoster, and while everything went well, RealRoster still making lots of NullReferenceExceptions in the log. (Also, the reason that RealRoster didn't seem to work last time is because I was only loading the Kerbal X. Even without the toolbar, it would correctly randomize crews when I added a pod to a vessel. It just doesn't work when loading a saved vessel, which is a known limitation.) I'll let enneract know about this. I doubt it has anything to do with the problem I first encountered, but bugs are bugs and need to be fixed, right?

Anyway, my next test will be a clean save with all of my mods, to see if I can make the issue happen again under similar circumstances to the first time (but only on a clean save). It may be memory pressure that's causing trouble, and the only way to get that is to put the mods back.

Edit: just did it. Only one NRE from RealRoster, a bunch of NREs from Clouds, no problems at all with switching vessels or the like. Next, I'm going to try running (a copy of) my main save where the problem happened the first time, to see if it still happens.

Edited November 5, 2014 by Kerbas_ad_astra
Results from testing day 2.

[diomedea]

I see you spent a lot of time writing about this, but really seems like you need no advice nor help, as you're proceeding to isolate the issue yourself effectively.

Some add-ons are pretty permissive about letting NREs being logged instead of filtering cases, but generally because the author knows those NREs have no effect. Sure I don't like that, as it makes bug chasing all the more difficult. I use NREs (in particular repetitive ones that seem connected to other abnormal manifestations) as a guide in the bughunting.

Interesting to see if your savegame has anything making RealRoster (or other add-ons) screw up. In case, correctly you should let the author know, he may consider the issue worth his time. Even if what is creating the issue was put in the save by KSP or any other add-on.

[Kerbas_ad_astra]

Thanks -- must be my engineering education paying off. I actually just got done trying out the save where I had the issue in the first place (with all of my mods back in and everything), and it worked like a charm. Even the capsule that I wasn't able to switch from or recover was right on the ground where I left it, and it was recovered without incident. Part of me wishes that this bug would occur more reliably so it could get stomped, but it's nice that nothing got permanently broken. Enneract (author of RealRoster) just produced an update that resolved the NREs that occurred when it was run without Blizzy's toolbar, so that's good, too. I'll mention this issue in the Trajectories thread as well -- I can't be sure that it's a cause, but it seems like the only mod I have that could conceivably be involved in patched conics.

Also, I wish I'd thought to try to EVA or IVA when I had this issue (it just occurred to me). I'll be sure to recall that next time.

[Paul Kauphart]

Hello,

I've been having the same problems multiple times with different missions and craft. I have no idea how to reproduce it, as far as I'm concerned, it could be random.

[HighLogic]: =========================== Scene Change : From FLIGHT to SPACECENTER (Async) =====================

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

NullReferenceException: Object reference not set to an instance of an object
  at PatchedConicRenderer.OnSceneSwitch (GameScenes scn) [0x00000] in <filename unknown>:0 

  at EventData`1[GameScenes].Fire (GameScenes data) [0x00000] in <filename unknown>:0 

  at HighLogic.LoadScene (GameScenes scene) [0x00000] in <filename unknown>:0 

  at AltimeterSliderButtons.returnToSpaceCenter () [0x00000] in <filename unknown>:0 

  at ScreenSafeUIButton.OnMouseUp (Boolean rightClick) [0x00000] in <filename unknown>:0 

  at ScreenSafeUIButton.OnMouseUp () [0x00000] in <filename unknown>:0 

  at SSUICamera.DoMouseEvents () [0x00000] in <filename unknown>:0 

  at SSUICamera.Update () [0x00000] in <filename unknown>:0 

(Filename:  Line: -1)

Game Paused!

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Flight State Captured

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Saving Achievements Tree...

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Saving Achievements Tree...

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Saving Achievements Tree...

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Saving Achievements Tree...

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Saving Achievements Tree...

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

Game State Saved as persistent

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

[HighLogic]: =========================== Scene Change : From FLIGHT to SPACECENTER (Async) =====================

(Filename: C:/BuildAgent/work/d63dfc6385190b60/artifacts/StandalonePlayerGenerated/UnityEngineDebug.cpp Line: 49)

NullReferenceException: Object reference not set to an instance of an object
  at PatchedConicRenderer.OnSceneSwitch (GameScenes scn) [0x00000] in <filename unknown>:0 

  at EventData`1[GameScenes].Fire (GameScenes data) [0x00000] in <filename unknown>:0 

  at HighLogic.LoadScene (GameScenes scene) [0x00000] in <filename unknown>:0 

  at PauseMenu.saveAndExit (GameScenes sceneToLoad, .Game stateToSave) [0x00000] in <filename unknown>:0 

  at PauseMenu.draw (Int32 id) [0x00000] in <filename unknown>:0 

  at UnityEngine.GUILayout+LayoutedWindow.DoWindow (Int32 windowID) [0x00000] in <filename unknown>:0 

  at UnityEngine.GUI.CallWindowDelegate (UnityEngine.WindowFunction func, Int32 id, UnityEngine.GUISkin _skin, Int32 forceRect, Single width, Single height, UnityEngine.GUIStyle style) [0x00000] in <filename unknown>:0 

(Filename:  Line: -1)

Game Unpaused!

I'm using KSP Windows Steam, 32bits, version 0.25.0.642. Current game has only a few mods :

DockingPortAlignment_4.0

ModuleManager-2.5.1

SCANsat-v8.0

Toolbar-1.7.7

The complete log is available here :https://www.dropbox.com/s/6dxfrwz74mhj0jy/output_log.txt?dl=0

[Paul Kauphart]

PS : For the record, my current workaround is : wait for the autosave, quicksave (just to be sure, never actually needed it), Alt-F4 the game and reload, problem gone.

[Kerbas_ad_astra]

Youen explained that Trajectories didn't use PatchedConicRenderer at all, and clearly you've experienced this bug without that mod being installed. I had wondered if maybe SCANsat was using it to draw orbit tracks on the big map, but a search of their source code shows only that they use patchedConicSolver (presumably to get the data which is then transformed to handle the rotation of the parent body and then drawn on the map).

[somewhatcasual]

I have this issue after using eva parachutes on a decently modded save. Were you able to sort it out, short of reloading ksp?

[Kerbas_ad_astra]

What you see is what we know. At least it doesn't seem to permanently break the save file. I've heard that Haystack Continued can force a switch between vessels when KSP won't do it, but I've never had this crash occur again so I can't test that.

[Djohaal]

1. [EXC 14:18:01.171] NullReferenceException: Object reference not set to an instance of an object
   PatchedConicRenderer.OnSceneSwitch (GameScenes scn)
   EventData`1[GameScenes].Fire (GameScenes data)
   HighLogic.LoadScene (GameScenes scene)
   PauseMenu.saveAndExit (GameScenes sceneToLoad, .Game stateToSave)
   PauseMenu.draw (Int32 id)
   UnityEngine.GUILayout+LayoutedWindow.DoWindow (Int32 windowID)
   UnityEngine.GUI.CallWindowDelegate (UnityEngine.WindowFunction func, Int32 id, UnityEngine.GUISkin _skin, Int32 forceRect, Single width, Single height, UnityEngine.GUIStyle style)
   

I am also encountering this exact same bug on a heavily modded 0.90 KSP. Occourence is random but I got a specific scene (a base in the mun) that always triggers it