[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[ferram4]

No, it just renamed the class.  And all the references to the class, so nothing broke.  It's just a silly name change for now.

[eberkain]

@ferram4 Struggling with a problem here, updated to the new version in my career game.  When I launch a vessel I get spammed with this which really lags the game down to just a couple FPS.  I try the same vessel in a new sandbox game and it gives no errors...   What do you think may have gone wrong and how can I fix it?

(Filename: C:/buildslave/unity/build/artifacts/generated/common/runtime/UnityEngineDebugBindings.gen.cpp Line: 64)

NullReferenceException: Object reference not set to an instance of an object
  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.GetVisibleMeshData (UnityEngine.Transform t, Boolean onlyMeshes) [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.CreateMeshListFromTransforms (System.Collections.Generic.List`1& meshTransforms) [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.RebuildAllMeshData () [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.FixedUpdate () [0x00000] in <filename unknown>:0 
 
(Filename:  Line: -1)

full output.log here https://drive.google.com/file/d/0B_wAo_KZdHCrU3g5Q0dsY28tdkU/view?usp=sharing

[ferram4]

@eberkain: Well, I don't get that when I launch a vessel, so that means that you're missing some crucial reproduction steps.  Narrow it down to the smallest number of steps possible with the smallest number of mods installed and I'll look at it.

[eberkain]

3 minutes ago, ferram4 said:

@eberkain: Well, I don't get that when I launch a vessel, so that means that you're missing some crucial reproduction steps.  Narrow it down to the smallest number of steps possible with the smallest number of mods installed and I'll look at it.

No sooner than I posted it I figured it out.  It would go away after I got a few KM from KSC, and would do it even if I went EVA.   I had a couple pieces of SEP equipment sitting next to the runway that I had setup and never recovered.  I recovered those and problem is gone now. 

[ferram4]

@eberkain: Uh... so then could you provide complete and full reproduction steps so that I can cause the issue and fix it?

[BigNose]

After some testing all I can say is, the problem of crafts behaving like there is a slight asymmetry even though they are mirrored seems to be either gone or unnoticeable now. Before I noticed it several times although I never thought it was that much of a problem, flying with joystick and hatkey for trim makes it easy to get rid of any constant roll.

Lanchester runs smooth and so far no further problems found! Nice work Ferram!

Edited July 3, 2016 by BigNose

[eberkain]

7 hours ago, ferram4 said:

@eberkain: Uh... so then could you provide complete and full reproduction steps so that I can cause the issue and fix it?

Ok, so I went to a clean install and added these mods.  

Then I started a new sandbox game and added a KIS container to the back of the stock vehicle crater crawler, loaded it with some SEP parts and drove off to the side of the runway and built a simple setup.  

I then returned to the space center and went back to the SPH and launched an aircraft, tried to take off and then checked the output_log and sure enough, getting spammed with these errors.  

NullReferenceException: Object reference not set to an instance of an object
  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.GetVisibleMeshData (UnityEngine.Transform t, Boolean onlyMeshes) [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.CreateMeshListFromTransforms (System.Collections.Generic.List`1& meshTransforms) [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.RebuildAllMeshData () [0x00000] in <filename unknown>:0 

  at FerramAerospaceResearch.FARPartGeometry.GeometryPartModule.FixedUpdate () [0x00000] in <filename unknown>:0 
 
(Filename:  Line: -1)

log file is here...

https://drive.google.com/file/d/0B_wAo_KZdHCrS2RQV3RvYlVJLVE/view?usp=sharing

It may not even be related to KIS / KAS / SEP, you might get this with anything that is surfaced attached, more testing is needed to narrow it down, but I hope this helps. 

[ferram4]

@eberkain: Alright, I'll look into it.  Thanks for the repro steps, they're exactly as detailed as I need them.  Should be easy enough to reproduce and fix when I get the free time to do so.

[Blocky999]

I appear to be having a problem.

As has been said previously (I think, from my quick glances at previous people's post), the fix is removing an engine.  In this case, it was the FS1PRO Customizable Propeller Engine from Firespitter.

Granted, my game is loaded up with 8+ mods, but none of the mods I have really seem like they would have an effect like this.

Edited July 7, 2016 by Blocky999

[NathanKell]

That is, shocker of shockers, a procedural part, ferram's nemesis... :\

[cantab]

Wow, as voxelization errors go, that one takes the biscuit. Unless it's remembering the voxels of what you had loaded before, I can't imagine what would cause such an error.

You also appear to have Tweakscale involved. I still recommend caution when using that with FAR.

[autumnalequinox]

On 7/4/2016 at 0:02 PM, ferram4 said:

@eberkain: Alright, I'll look into it.  Thanks for the repro steps, they're exactly as detailed as I need them.  Should be easy enough to reproduce and fix when I get the free time to do so.

I got the same bug too with KIS/KAs and the SEP stuff.  Only when changing flight scene to Minmus base consisting of two linked craft and a SEP setup.  I deleted the SEP setup (through terminate) and it reduced the lag but still threw up same exceptions.  Just confirming the bug.

[Blocky999]

12 hours ago, cantab said:

Wow, as voxelization errors go, that one takes the biscuit. Unless it's remembering the voxels of what you had loaded before, I can't imagine what would cause such an error.

You also appear to have Tweakscale involved. I still recommend caution when using that with FAR.

I did indeed have another ship loaded prior.  As for Tweakscale, I have yet to have issues with it and FAR when scaling default parts.  Can't say the same for any mods because I don't typically use tweakscale for anything besides wheels and wings, both of which are always the defaults.  Also, in addition to the voxel issue, I was also not able to get the stability derivatives to recalculate when I loaded the ship with that procedural engine.  Although, that was when I hadn't loaded another ship prior, so I would expect it to show the derivatives for the previously loaded ship like it shows the voxels for that previous ship had I checked the derivatives when I snapped that first pic.

[Vegetal]

@ferram4 bad news. I just tested that dll you sent and then had to stay away from KSP for a while. Now I'm back, downloaded Lanchester and the bug persists. Voxels seem almost perfectly symmetric, but the plane behaves just like the first time I reported.

I don't even know what to report man.

[123nick]

hey the FAR abbreviation on the forums, when moused over, says : 

aeronautics instead of aerospace, as the title of the mod actually is.

[kcs123]

7 hours ago, 123nick said:

hey the FAR abbreviation on the forums, when moused over, says : 

aeronautics instead of aerospace, as the title of the mod actually is.

That should be adressed to @KasperVld, I belive. There was a discussion about it in commonly used acronyms thread. Probably just a typo due to lack of proper sleep and too much of tedious work.

[Paul_Sawyer]

Looks like the voxel problem is back indeed. I have this craft with the whole right side full of dots sticking out. It also tends to deviate to the right while flying

 

 

If you need the craft it is here.

Edited July 18, 2016 by Paul_Sawyer

[blowfish]

Hmm, I recall the collider having normal issues on that particular part.  It's possible Squad has fixed it since though.

[Paul_Sawyer]

@blowfish But it's not only this part:

[tetryds]

@Paul_Sawyer As long as that does not affect flight performance I wouldn't be concerned.

I have made airplanes that fly at Mach 2.7 bellow 500m and never had any problems whatsoever, are you experiencing any?

Remember that FAR doesn't use pure voxels, there are voxel halves, corners and stuff like that (AFAIK), but it is indeed weird that you have big voxels on the left.

Does this happen to every airplane?

[Paul_Sawyer]

1 hour ago, tetryds said:

@Paul_Sawyer As long as that does not affect flight performance I wouldn't be concerned.

I have made airplanes that fly at Mach 2.7 bellow 500m and never had any problems whatsoever, are you experiencing any?

Remember that FAR doesn't use pure voxels, there are voxel halves, corners and stuff like that (AFAIK), but it is indeed weird that you have big voxels on the left.

Does this happen to every airplane?

Well as I said, the first craft flies really badly, it tends to roll right every time I pull it up. Although it is a typical illness for my planes, but usually it is manageable and for this one it is deadly.
As for voxels, apparently it depends on part. Up to now I found problems with mk3 adapter, mk2 small cargo bay and bicoupler, mk2 fuel fuselages, also LY-01 landing gear has this problem when placed symmetrical.

Edited July 18, 2016 by Paul_Sawyer

[kcs123]

Silly question, but have you tried without FAR flight assistance(roll/yaw/AoA) ?

I was noticing asymmetric voxelization in previous FAR/KSP releases, but it were not giving me any noticable issues in flight, so I didn't even bothered to report.
Don't even recall when i was last time checked voxelization, there was no need for it at all.

However, I was noticing issues with FAR flight assistance, more often when you need to quicksave/quickload. Wasn't bothered me too much, I rarely use those, so I didn't put notes on reliable reproducing steps and to report it here either.

Might be wrong, but more dense or rare voxels on one side or other should not have influence if on both sides mesh/collision is considered as solid closed surface. Voxelization does not add any asymetric mass on parts either that could have significant impact on unstability.

[blowfish]

1 hour ago, Paul_Sawyer said:

@blowfish But it's not only this part:

Just checked, and both the Long Mk2 fueslage and the bi-coupler seem to have mesh errors in their colliders.  In the first instance, there are zero-area faces.  In the second, the normals for one entire side of the bi-coupler are inverted.

[DownHereInChile]

So it's been a little while since I last played ksp, and I noticed that the CoL in my aircraft moved drastically to the front with the latest FAR patches, turning my once very stable and maneuverable planes into death-tubs. Has anyone else encountered this problem? Any clues as to why it may be happening? Has this been previously answered?

[Qwits]

Does this mod have a possibility of independently tweaking the drag? There's an option called "Frac Mach 1 Drag", but changing it from 0 to 100000 doesn't seem to change anything, neither do the presets (what's its purpose?). Tweaking viscosity does work, but to reach stock-alike drag levels requires cranking it to unreasonable values. I'd appreciate some help on this as at the moment I'm opting for stock aerodynamics as giving me the freedom to set it up to my liking.

Edited July 22, 2016 by Qwits