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

[Mighty1]

On 5/25/2017 at 8:07 AM, ferram4 said:

Get the most specific reproduction steps you can, ideally ones that create the issue reliably all the time (if you can't, most of the time) and then post a copy of the save file that causes this.  Post it all on Github so it's easier to track and I'll look into it.

Probably something coincidental though.  Most crashes are memory related and I'm not sure how this could result in some memory access or memory allocation error.

Post a copy of the craft file and make a Github issue for it.  In addition, if other mods are required for this craft, simplify it down to the point where only the mods absolutely necessary to cause the issue are involved and then list every single mod needed.  If you don't do this, I can't guarantee that I can reproduce the issue considering how variable designs can be.

 

Issue reported on GitHub

[Citizen247]

Congrats on the release. I've been testing and noticed that with AJE engines seem to cause FAR to constantly revoxilize. I've only played around a bit, but everytime I place a craft on the runway with an AJE configured Jet I get performance issues and revoxlizing messages in the console. Same craft minus the engine and I get no messages. Is this likely to be a far issue or an AJE issue?

[LOG 00:17:03.170] Updating vessel voxel for FAR-ieee
[LOG 00:17:03.237] Voxel Element CrossSection Area: 0.00924555692500613
Std dev for smoothing: 3 voxel total vol: 222.248901981287 filled vol: 12.7334084139059

[LOG 00:17:04.185] Unpacking FAR-ieee
[LOG 00:17:04.859] Updating vessel voxel for FAR-ieee
[LOG 00:17:04.918] Voxel Element CrossSection Area: 0.00924601218369

[LOG 00:17:04.946] Std dev for smoothing: 3 voxel total vol: 222.265317756287 filled vol: 12.7285682342841

[LOG 00:17:05.189] [Progress Node Reached]: Landing
[LOG 00:17:05.189] [Progress Node Complete]: Landing
[LOG 00:17:06.565] Updating vessel voxel for FAR-ieee
[LOG 00:17:06.616] Voxel Element CrossSection Area: 0.00924592113284988

[LOG 00:17:06.645] Std dev for smoothing: 3 voxel total vol: 222.262034601287 filled vol: 12.7291872165651

[LOG 00:17:08.245] Updating vessel voxel for FAR-ieee
[LOG 00:17:08.308] Voxel Element CrossSection Area: 0.00924596418795419
Std dev for smoothing: 3 voxel total vol: 222.263587100938 filled vol: 12.7293298353944

[LOG 00:17:09.966] Updating vessel voxel for FAR-ieee
[LOG 00:17:10.016] Voxel Element CrossSection Area: 0.00924596559959525

[LOG 00:17:10.067] Std dev for smoothing: 3 voxel total vol: 222.263638002566 filled vol: 12.7284065205072

[LOG 00:17:11.654] Updating vessel voxel for FAR-ieee
[LOG 00:17:11.733] Voxel Element CrossSection Area: 0.00924596524668499
Std dev for smoothing: 3 voxel total vol: 222.263625277159 filled vol: 12.7284023052713

[LOG 00:17:13.320] Updating vessel voxel for FAR-ieee
[LOG 00:17:13.383] Voxel Element CrossSection Area: 0.00924596454086446
Std dev for smoothing: 3 voxel total vol: 222.263599826345 filled vol: 12.7284373670297

[LOG 00:17:15.028] Updating vessel voxel for FAR-ieee
[LOG 00:17:15.094] Voxel Element CrossSection Area: 0.00924599242075504
Std dev for smoothing: 3 voxel total vol: 222.264605133496 filled vol: 12.7285142302196

[LOG 00:17:16.727] Updating vessel voxel for FAR-ieee
[LOG 00:17:16.774] Voxel Element CrossSection Area: 0.00924600583132035

[LOG 00:17:16.822] Std dev for smoothing: 3 voxel total vol: 222.265088698961 filled vol: 12.7303096112168

[LOG 00:17:17.567] [Kopernicus] New object found near Ciro: Ast. NFK-889!
[LOG 00:17:17.567] [Kopernicus] No new objects this time. (Probablility is 35%)
[LOG 00:17:17.586] FARVesselAero on Ast. NFK-889 (unloaded) reporting startup
[LOG 00:17:18.467] Updating vessel voxel for FAR-ieee
[LOG 00:17:18.525] Voxel Element CrossSection Area: 0.00924593595511022

[LOG 00:17:18.576] Std dev for smoothing: 3 voxel total vol: 222.26256906838 filled vol: 12.7302306941272

[LOG 00:17:20.159] Updating vessel voxel for FAR-ieee
[LOG 00:17:20.221] Voxel Element CrossSection Area: 0.00924597618689982
Std dev for smoothing: 3 voxel total vol: 222.264019764775 filled vol: 12.7284652714464

[LOG 00:17:21.858] Updating vessel voxel for FAR-ieee
[LOG 00:17:21.935] Voxel Element CrossSection Area: 0.00924598183345977
Std dev for smoothing: 3 voxel total vol: 222.264223371287 filled vol: 12.728521585983

 

[blowfish]

@Citizen247 I probably need to add AJE's animation modules to the FAR voxel overrides.

[Citizen247]

9 hours ago, blowfish said:

@Citizen247 I probably need to add AJE's animation modules to the FAR voxel overrides.

Yep, I added:
 

FARAnimOverride
{
	moduleName = ModuleAJEJetAnimateNozzleArea
	animNameField = animationName
}
FARAnimOverride
{
	moduleName = ModuleAJEJetAnimateAfterburner
	animNameField = animationName
}
FARAnimOverride
{
	moduleName = ModuleAJEJetAnimateExhaustHeat
	animNameField = animationName
}

Into FARAnimOverrides.cfg and the revoxilisation has vanished. Good call thanks.

[meatshield_chris]

On 5/26/2017 at 9:14 PM, meatshield_chris said:

I just installed the latest FAR on a fresh 1.3 install alone, no other mods. The aerodynamics tweaks appear to work, but the interface doesn't show. Might be an easy fix. 

I was wrong about that, looks like drag being mostly ignored on a capsule re-entry that I can tell. (Just a pod and a parachute, no heat shield or other things on it) The lift changes still work, you can still kinda fly the pod a bit, affecting the orbit while in atmosphere, but not slowing down. Tried lots of different methods of slowing down, but even deep in the atmosphere it would accelerate like the atmosphere wasn't there if you were moving in the downwards direction. A short hop leaning way over so that you stay in the atmosphere keeps accelerating all the way to the ground. If I strap airbrakes to it, I can get it slowed down. Makes the early career mode impossible though. 

So I guess it's not quite fully working aerodynamically on 1.3.

[ss8913]

I can't get KSP to even start with any module not specifically compiled for 1.3 loaded in - FAR included.  I have a "working" game but as soon as I put in FAR (or any other non-1.3-specifically built module), the entire game crashes on load at ModuleManager:FINAL

I'll try the latest dev build when I get back from the store.

[ferram4]

Stop trying to make FAR work on 1.3.  I'm still looking at some development tweaks in 1.2.2, possibly another patch release shortly so that RO doesn't have the splashed-down unpacking crash bug that was mentioned earlier.  I haven't even looked at a 1.3-compatible MFI yet.

Actually, I haven't even downloaded 1.3 yet.

[ss8913]

17 minutes ago, ferram4 said:

Stop trying to make FAR work on 1.3.  I'm still looking at some development tweaks in 1.2.2, possibly another patch release shortly so that RO doesn't have the splashed-down unpacking crash bug that was mentioned earlier.  I haven't even looked at a 1.3-compatible MFI yet.

Actually, I haven't even downloaded 1.3 yet.

fair enough; this saves me some time trying out various branches/etc   will wait for a 1.3 release or a 1.3 dev build.

[Calvin_Maclure]

31 minutes ago, ferram4 said:

Stop trying to make FAR work on 1.3.  I'm still looking at some development tweaks in 1.2.2, possibly another patch release shortly so that RO doesn't have the splashed-down unpacking crash bug that was mentioned earlier.  I haven't even looked at a 1.3-compatible MFI yet.

Actually, I haven't even downloaded 1.3 yet.

I'm still on 1.1.2! lol... 

[asaz989]

tl;dr - craft launched before installing FAR have zero/NaN lift/drag coefficients. Is there any way to force a recalculation of craft shape (revoxelization?) in-flight?

I have an existing save that I put on hold while waiting for FAR to be ready for 1.2.2. Specifically, I stopped playing around when I started needing to do serious aerobraking work at Jool. Now that I have installed the new FAR version, my probes hitting Jool are not being braked at all by the Joolean atmosphere. The flight data window shows zero/NaN coefficients of lift/drag, zero reference area, and zero drag/lift forces. (They *are* being properly ripped apart by dynamic pressure though, which is something .) Newly-launched craft are properly interacting with Kerbin's atmosphere, though.

I vaguely remember seeing something like this mentioned a while ago for a previous release, but my search-fu is not sufficient to figure out if that's just my mind playing tricks on me.

Edited May 29, 2017 by asaz989

[ferram4]

Revoxelization always happens once the craft is loaded, and as far as FAR is concerned, there is no difference between a craft that has just been loaded and was already existing in a save and one that was just launched.

Make an issue on github to make it easy to track and include the save file and the craft file responsible.

[asaz989]

Awesome. Will do when back at computer.

Edited May 29, 2017 by asaz989

[asaz989]

Okay, false alarm. After cleaning up weird mod versions for a neater bug report, the problem went away.

In case you have a morbid curiousity about weird installs - at one point I let Steam auto-upgrade to 1.3.x, ran CKAN, and then realized I should downgrade to 1.2.2 so I'd have my mods available. Anyway, several mods were left at versions that supported 1.3.x but were incompatible with 1.2.2, including ModuleManager, and CKAN will not downgrade packages through the GUI, only when run at the command line. Other mods were also broken-ish, but ModuleManager seems the scariest and the most likely to have caused bugs.

Edited May 29, 2017 by asaz989

[John FX]

Myself I am glad that FAR has caught up with 1.2.2, another step done on the path to RSS/RO/RP-0 in 1.2.2. I also have not even downloaded 1.3 yet.

Thank you for your time and effort.

[Moach]

On 5/23/2017 at 1:27 AM, Mighty1 said:

I reported this issue a while ago and thought that it was a dev version problem but it is happening again.

So, a clean install, only Modular Flight Integrator 1.2.4,ModuleManager 2.7.6 and FAR De Laval installed. I landed a stock Gull in the water some 15 km off the KSC in the water, launch another gull and approach. When I get to cca 200 m distance the game crashes. Sometimes there is no crash but the targeted vessel just vanishes and lands on the sun accoring to the tracking station. I tried the same approach while landed and nothing happens and in orbit there are no issues also. It appears it is a water thing.

Here is the crash log https://www.dropbox.com/s/ikrh70zh7sxyfdi/2017-05-23_093008.rar?dl=0

 

 

I would like to confirm this report, as I have also encountered it 

 

it was noticed on an attempted rescue mission for a botched orbiting - that is, the downed vessels were left floating (splashed down) and the rescue airship was dispatched

upon the airship nearing 200 meters off the downed vessel (entered physics range) - it was like a delete-brush effect which caused everything that entered it's insta-doom radius to instantly dematerialize 

 

the logs revealed a series of NAN values following the "recalculated orbit for..." entry immediately before the madness of null-exceptions and consequential symptoms that ensued

curiously, the preceding "recalculating orbit..." entry appeared fine (numbers, where one would expect them) - but after that, all hell broke loose

 

 

and my crews all died... I mean, died slightly sooner than expected, since they we're slated to be sent up again

[Nothalogh]

On 5/30/2017 at 7:16 PM, Moach said:

my crews all died... I mean, died slightly sooner than expected, since they we're slated to be sent up again

The managerial mindset

[AdmiralTigerclaw]

Hey Ferram, is revoxelization supposed to occur constantly in flight?

 

I ran a flight yesterday that was supposed to take 4 hours round trip.  However, about 2 hours in, the flight physics slowed to a crawl (33%) of realtime, and I started noting in the debug menu that the flight was revoxelizing every few seconds.  The flight ended up taking 7 hours to complete.

I'm trying to narrow down my culprit for processing time consumption since it seems that while in the map, time was normal, but in normal camera view, it was... As I pointed out, a crawl.  Long haul atmospheric flights always seem slow down for me given enough flight time.  Did it since at least 1.0 if not before.  But I noted the voxelizing constantly this time around, and thought I'd drop it in for commentary

 

The craft in this was using B9's basic wing jigsaw pieces. 

 

Here's a chunk from the game log showing the voxel updates.  They were frequent.  (And sheesh, the decimal places kept for that...)

Quote

[LOG 14:27:22.345] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:22.447] Voxel Element CrossSection Area: 0.0237615142568321
Std dev for smoothing: 1.41584927466441 voxel total vol: 915.695709017194 filled vol: 17.2864947390041

[LOG 14:27:27.628] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:27.729] Voxel Element CrossSection Area: 0.023761528127625
Std dev for smoothing: 1.41620777528175 voxel total vol: 915.696510823881 filled vol: 17.290886912362

[LOG 14:27:31.070] [WBIDistributionManager] - Using distributor cache
[LOG 14:27:32.873] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:32.979] Voxel Element CrossSection Area: 0.023761528127625
Std dev for smoothing: 1.41612790908628 voxel total vol: 915.696510823881 filled vol: 17.289911803075

[LOG 14:27:38.137] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:38.302] Voxel Element CrossSection Area: 0.0237615142568321
Std dev for smoothing: 1.41633541076119 voxel total vol: 915.695709017194 filled vol: 17.2924301088416

[LOG 14:27:43.106] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:43.194] Voxel Element CrossSection Area: 0.0237615038537349
Std dev for smoothing: 1.41623497246178 voxel total vol: 915.69510766218 filled vol: 17.2911924744445

[LOG 14:27:47.976] Updating vessel voxel for OX-A1 Scout
[LOG 14:27:48.065] Voxel Element CrossSection Area: 0.0237615399177957
Std dev for smoothing: 1.41635925857714 voxel total vol: 915.697192359564 filled vol: 17.2927492860208

 

 

Looks like, from here, that it was constantly updating what I'm assuming to be fuel volume as it's consumed, since the numbers decrease at a crawl.

[Jack_Aubrey]

I think I will download the 1.2.2 and wait to use 1.3 .... the thing is that some mods are a must've ... and one is FAR. THX Ferram !!!

[Guest]

On 5/28/2017 at 6:12 PM, Calvin_Maclure said:

I'm still on 1.1.2! lol... 

I was... until 2 months ago.

Oh, and thanks Ferram4, ive been waiting for the 1.2.x release.

Edited June 1, 2017 by Guest

[Kussris]

On 22.5.2017 at 8:31 PM, ferram4 said:

Oh, also, FAR v0.15.8 "de Laval" is out now.  That name may be familiar to you....

 

Thank you ferram4 for updating FAR to KSP 1.2.2 !!!!!!!!!!

 

[KaiserSoze]

What are the affections of 1.3 to this mod?

It was my undertanding that the main changes were in the ui, so that is the part affected? or it is something else hidden that is also disturbing it?

 

Please don't misundertand me, I do not want to sound like I want this mod to be updated now (well, I do, but I undertand it takes works and that people that have life outside KSP) I just want to figure what part of the mods need to be worked on (again)

[ss8913]

8 hours ago, KaiserSoze said:

What are the affections of 1.3 to this mod?

It was my undertanding that the main changes were in the ui, so that is the part affected? or it is something else hidden that is also disturbing it?

 

Please don't misundertand me, I do not want to sound like I want this mod to be updated now (well, I do, but I undertand it takes works and that people that have life outside KSP) I just want to figure what part of the mods need to be worked on (again)

I tried using the current version in 1.3 - crashes on load, will not start at all.  It at the very least needs to be recompiled for 1.3 (as well as modular flight integrator).  Not sure if anything else needs updating.  Have seen some mods having font trouble in 1.3, (mods that were built for 1.2.x) so, that may or may not need to be addressed in FAR.  I haven't attempted a recompile myself.

 

[blowfish]

It has been my observation that plugins require recompiling for KSP 1.3 even when no code changes are necessary.  I wouldn't try to pull anything forward except for parts-only mods.

[taniwha]

As far as I can tell, ModularFlightIntegrator from 1.2 works in 1.3: I got FAR working locally last night without recompiling MFI (FAR needed some api fixes).

I suspect it's mods with PAW events that cause problems due to that particular API change (don't quote me on this as I'm merely pulling rabbits out of hats here, but that might explain the need for a recompile even without api fixes).

[ss8913]

3 hours ago, taniwha said:

As far as I can tell, ModularFlightIntegrator from 1.2 works in 1.3: I got FAR working locally last night without recompiling MFI (FAR needed some api fixes).

I suspect it's mods with PAW events that cause problems due to that particular API change (don't quote me on this as I'm merely pulling rabbits out of hats here, but that might explain the need for a recompile even without api fixes).

so I could potentially get FAR working by firing it up in visual studio, editing the include paths in the project file, and building a new DLL... keeping MFI from 1.2.2 the same?  I may do so.