[1.3] Kerbal Joint Reinforcement v3.3.3 7/24/17

[ModZero]

I put this in Sounding Rockets thread, because this only happened to me in that mod, but it's an interaction with KJR, so I'll cross-post:

Hi there, I come bearing bugs (possibly not in this mod, though - also sorry if this is known, I didn't spot it in a search).

If there's Kerbal Joint Reinforcement installed, first stage of a two-stage SR doesn't drop if there are fins attached to the first stage, but it works correctly on a rocket without these fins. To make things even more confusing, this (sounding rocket with fins, but using small boosters for both stages) works properly.

Sorry if it's not your bug - it seems to be an interaction with KJR. Here's a log of a single launch of the broken rocket.

EDIT: also, I remember it happening before I upgraded KJR, but I didn't do any isolated tests back then.

[Galane]

I need full reproduction steps and output logs in order to address any issues. I have not sen any issues with regards to clamps or decouplers, and nothing has changed with how they are handled at all.

There is some stock issue with the 1-to-n adapters, especially when stages are attached radially to tanks attached to the bottom sides of the adapters, or if the adapters are used upside down. I first notices that with this project in the version with the four missiles. Switch around the order of the four missile stages and the decouplers will not trigger.

I built an all stock parts rocket in .25 with the 1-to-4 adapter then used radial decouplers to mount Kerbodyne tanks to the sides of the tanks attached to the bottom of the decoupler. Those side boosters had second tanks atop them, with the big decouplers and plenty of Sepratrons to kick them away. With the decouplers right side up it worked. Flip them upside down to drop their weight with the tanks and it all goes worng.

The same node order issue appears, not because of staging individually, but because the Kerbodyne size decouplers have Physics Signifigance set to 1 so they have no mass in flight. (Cheaty by default!) The problem also manifests when using an up-scaled version of the 1-to-n adapters to fit Kerbodyne tanks. With Physics Significance left at 1, those decouplers directly in line beneath those adapters will not all activate. One will, the next stage engines will, but the other 1, 2 or 3 big decouplers fail. *big explosion* as the engine flames destroy the decouplers.

So if you are having staging problems using the kerbodyne decoupler in anything but a single stack rocket or in an asparagus rocket where Kerbodyne size stacks are not attached below a stock 1-to-n adapter, you need to enable Physics Significance in the Kerbodyne decoupler.

Squad has kept that error in .90 and I don't know why. Those decouplers are supposed to be heavy parts that require appropriate rocket design to account for the mass, unlike struts and various other small parts. None of the other decouplers or stack separators has that setting and they cause no issues in any configuration I've used them in.

As long as you're not trying to do sequential staging off the bottom of a 1-to-n adapter (right side up or otherwise) there shouldn't be a problem - unless by some weird thing the latest KJR is interacting with the bug in how the multi-node side of those stock adapters operates?

[ferram4]

@ModZero: Nah, that's on my side; I see what's happening. OnPartJointBreak doesn't fire for decoupling events, and since USI drop tanks and sounding rockets go around that, KJR doesn't clean up like it should. Dev version should have a fix.

@Galane: Hmm... I just tried playing around with that, and I wasn't able to reproduce it. Could you throw together a small test vehicle for me to mess with? I'm slightly concerned that I'm missing something important in throwing it together.

Edit: You know, just once I'd like an x.0 release to be bug-free.

[NuAoA]

I just updated this mod along with Kerbal engineer redux, and occasionally when I drop out of a time warp too fast my ship will combust or my velocity will drop to ~0 and become unchangeable. I've progressed pretty far through career, including several landings with the ship i'm using and never had this issue until I updated.

Edit: Removed KJR and the issue was gone. It was happening on a sub 6000m minmus orbit, and didn't happen until I got close to the planet. I've also switched back to 2.4.5 and the issue is gone.

Edited December 29, 2014 by NuAoA

[NuAoA]

Is there a 2.4.5 version kicking around? Its not on the earlier versions list.

Edit: its on kerbalstuff, don't ask me how I found it: [Defunct link removed by a moderator]

Edited August 28, 2022 by James Kerman
link removed

[ferram4]

@NuAoA: I cannot reproduce the issue following the little information you have provided. I will need full reproduction steps, a full copy of the output log, and a full list of all mods that you have been running in order to address the issue. Less than that leaves out critical information and makes fixing the issue impossible. Why do people never provide logs?

[Voidryder]

Why do people never provide logs?

Just for you ferram4.

http://nicktoons.nick.com/videos/clip/stimpys-big-day-log-song-1.html

[NuAoA]

Most likely because the average person doesn't mod ksp or have a clue about what logs you are talking about.

All of these mods plus KW Rocketry: http://i.imgur.com/qlWBhxy.png

Zip of my save: http://www.filedropper.com/tothemun. KSP saves compress surprisingly well (~8%!).

The ship is called something like mun lander mk2 and is orbiting minmus.

[ferram4]

The save doesn't help me, I'm not able to derive what went from from that, nor am I sure that I can get the exact install that you have there (maybe some things aren't updated, who knows).

Follow this

[NuAoA]

Heres the log, but as a heads up, I never saved the game after it happened and i've revered back to 2.4.5 and continued on since.

http://www.filedropper.com/outputlog_5

As to how to recreate it, I gave you as much detail as I could. The ship does involve some oddly placed and/or overlapping parts, so if you could get it loaded from the save (Its saved as a VAB ship in there, 'mun lander mk2') that would be your best chance at recreating it. Everything went fine with the launch/transfer to minimus, and once I got close to the planet (~6000m periapsis) it would blow when I dropped out of timewarp.

Every mod on that list is the newest version, except realchute and KAC, which are probably 1 version behind.

Edit: the only other tidbit was that the first collision occurred on a decoupler, the up13 decoupler from deadly reentry I think, otherwise it was the stock tt-38k. And I only checked what caused it once, so those might just be random parts.

Edited December 29, 2014 by NuAoA

[Galane]

Here's an all stock test craft demonstrating the sequential staging bug in the stock 1-to-4 adapter. http://pastebin.com/dA1Prjny Saved in .25, tested in .25 and .90 to verify the bug in both versions.

Swap the positions of any two stages in the stack in the VAB and all four will fire correctly. Leave it as it is and the second stage will not activate the engine. It doesn't affect Sepratrons, just liquid fuel engines. I just tested it many times in both KSP versions to verify that swapping *any two* stage stack positions will have the probes launching in the assigned order while leaving it as-built causes the second probe's engine to fail to activate.

Why? I dunno. Bugs tend to be that way. Tricky bastards. Been there since at least .21, which is the first version I used and when I noticed this problem. If it can be fixed by making changes to the cfg files for those adapters, then I'll apply that to the ReStock part. Fixing the original? That's up to Squad, or individuals to do on their own if a fix is found.

If you want, I can re-build a sure-fire-fail rocket with all stock parts, using Kerbodyne stacks hung off the sides of small stacks below this quad adapter. But it will only exhibit the failure if the Kerbodyne decoupler has the Physics Significance = 1 line left active. Should also be able to replicate the partial stage separation failure with the massless Kerbodybe separator used below the Bigger, Beastlier adapters. 'Course those issues are easily fixed by removing the physics line.

[mike9606]

Well, this seems to be broken. Trying to launch a tantares station from the Kosmodrome (kerbal konstructs thing) pad using the ALV launch vehicle, I could not use launch clamps as when staged they acted as if detached in functionality, but did not animate and remained attached to the rocket. Flying without clamps, once I staged away Block A, the command module tore itself away from the ship. I have no idea why it is acting up, but I suspect it has something to do with the multipart joints for the command module tearing away, and it seems as if KJR isnt considering the clamps detached when I stage or something like that.

[ferram4]

@NuAoA: I haven't been able to reproduce it in the latest build, even while coming out of high timewarp at ridiculous velocities near Minmus. I suspect another mod is to blame.

@Galane: Ah, I see. Yay, stock bugs. I'm not exactly sure how to fix it, tbh.

@mike9606: No output log, no support. I have been unable to reproduce the issue on my build, so I will need full reproduction steps using as few mods as possible. If it is due to KJR, you should be able to cause it with no other mods.

Anyway, KJR v3.0.1 is out, fixing some of the new multipart joint shenanigans, particularly with USI drop tanks and sounding rockets.

[JeffreyCor]

Thought I'd give a heads up, it seems if you have a ship already launched updating this will cause it to rip itself apart or otherwise do really weird things like self propelling itself under torque somehow. Reverting back to 2.4.5 does correct it for continued play so landing / disposing of current flights before installing 3.0.1 seems to be advisable.

[ferram4]

Wait, all of the reports require already-existing ships? If that's the case, why didn't anyone else mention that? That could be really important!

But KJR doesn't save anything that could be reloaded anyway... so what?! How?

[JeffreyCor]

Weirdness seems to be KSP's stock in trade Don't know about anyone else, but 3 loads with 3.0.1 all caused existing ship sadness (crashed out to prevent saving) but reloading after reverting to 2.4.5 allows everything to continue as usual first try.

[WuphonsReach]

I haven't been able to triage fully... but weirdness for me involves the tiny cubic frame part or the tiny octagon shaped frame part. I've had a few rockets with those which would not lift off as one or more of those parts were welded to the universe (turned into immovable objects). Possibly only bugs out if I start stacking them.

But not 100% that it's KJR doing it, so might be a useless lead.

[Galane]

@WuphonsReach Do these craft by any chance happen to include any TR-38-D? NASAmission/parts/size3decoupler Open the CFG and insert // in front of the Physics Significance line or change the 1 to a 0. Your crafts using that will gain a bit of weight but the flipping weirdness sometimes caused by having such a large part "massless" in flight will go away. None of the other separators or decouplers have that setting, I dunno why Squad put that in the TR-38-D.

Yes, KSP seems to revel a bit in weirdness. When I was testing my assemble-on-site Gilly refueling lander it suddenly developed an aversion to having Quantum Struts aimed at some parts that had never given trouble before - and this was one day to the next in the same install, probably some funky interaction with one or more other mods recently updated.

Was a real hair puller figuring out why the thing suddenly was spinning out of control or wouldn't maneuver correctly and would rip itself apart when hyperedited to Gilly orbit. Took off two QStruts and the problem went away.

Still waiting on some mods to get sorted out for .90 so I've been working a bit on ReStock updates and fixes.

[King Arthur]

Wait, all of the reports require already-existing ships? If that's the case, why didn't anyone else mention that? That could be really important!

But KJR doesn't save anything that could be reloaded anyway... so what?! How?

This sounded easy enough to test and try reproducing, so I tried the following to see if KJR v3.0.1 would blow up a pre-existing stock Kerbal X:

1. Made a new blank sandbox save.

2. Launched a Kerbal X under pure stock, no mod (not even KJR) environment into a stable orbit.

3. Exited KSP and installed KJR.

4. Booted up KSP (with KJR installed) and loaded up the save with the Kerbal X.

5. Switched to it through the Tracking Station.

6. KJR did its magic.

7. Nothing out of the ordinary happened.

The end result was that KJR did not do anything bad or not intended to a pre-existing Kerbal X, in an environment where no other mods were at play.

The logs, even though I doubt they'll be interesting:

http://rpgmaker.net/media/content/users/6088/locker/PreKJR_output_log.txt

http://rpgmaker.net/media/content/users/6088/locker/PostKJR_output_log.txt

[JeffreyCor]

That did not test upgrading from KJR 2.4.5 to 3.0.1 so did not attempt to reproduce the problem.

[King Arthur]

That did not test upgrading from KJR 2.4.5 to 3.0.1 so did not attempt to reproduce the problem.

Ah, I missed that part. My bad, will go and test that (and edit this post once I get something post-worthy).

---EDIT---

Mod present: KJR 2.4.5 -> 3.0.1

Steps to reproduce:

1. Made a new blank sandbox save, KJR 2.4.5 installed.

2. Launched Kerbal X to orbit.

3. Exited, deleted KJR 2.4.5, installed KJR 3.0.1.

4. Booted up KSP after KJR 3.0.1 install, loaded save with Kerbal X in orbit.

5. Switched to Kerbal X via Tracking Station.

6. KJR did its magic.

7. Nothing out of the ordinary happened.

The end result was that, once again, KJR did not do anything bad or not intended to a pre-existing Kerbal X, in an environment where no other mods were at play.

The logs:

http://rpgmaker.net/media/content/users/6088/locker/245_output_log.txt

http://rpgmaker.net/media/content/users/6088/locker/301_output_log.txt

[ErrHead]

With the new version of KJR my rockets have every joint fail as soon as KJR initializes physics, resulting in every part falling down individually and exploding on the pad.

I run realism overhaul on KSP 0.90 linux x64 (on debian wheezy). The rockets in question do happen to be very large, my goal is usually around 1000 metric tons to LEO per launch, but everything worked and works just fine with KJR v2.4.5

I tried reinstalling KJR manually instead of through CKAN, nothing changed.

Output.log:

https://www.dropbox.com/s/q72myl45gtclma8/Player.log.zip?dl=0

Here is another log with KJR's debug option turned on:

https://www.dropbox.com/s/h64enuif1al45f3/Player.log2.zip?dl=0

Here is the same log, with debug on, but from version 2.4.5 where my rocket does not automatically disassemble.

https://www.dropbox.com/s/my657v4mart6srs/Player.log2.4.5.zip?dl=0

Gamedata:

GameData/
000_Toolbar
ActiveTextureManagement
AIES_Aerospace
AJE
AmbientLightAdjustment
blackheart
BobCatind
Chatterer
CollisionFX
CrossFeedEnabler
CustomBiomes
DDSLoader
DeadlyReentry
DistantObject
DMagic Orbital Science
EditorExtensions
EngineGroupContoller
EnvironmentalVisualEnhancements
FASA
FerramAerospaceResearch
Firespitter
FreedomTex
ImprovedChaseCamera
JSI
KAS
KerbalJointReinforcement
Klockheed_Martian
Klockheed_Martian_SmartParts
KSP-AVC
MechJeb2
ModuleRCSFX
NASAmission
NavyFish
NearFutureConstruction
NearFutureSolar
NEBULA
NothkeSerCom
PlanetShine
ProceduralFairings
ProceduralParts
RCSBuildAid
RealChute
RealFuels
RealismOverhaul
RealSolarSystem
RemoteTech
RLA_Stockalike
RN_R7
RN_Soviet_Probes
RSSTextures
RVE
SCANsat
SmokeScreen
Space Factory
Squad
StretchySNTextures
TextureReplacer
ThunderAerospace
TriggerTech
TweakScale
VNG


61 directories

AVC Output:
KSP: 0.90 (Unix) - Unity: 4.5.5f1 - OS: Linux 3.2 Debian 7.7 64bit
Ambient Light Adjustment - 1.3.1.1
Chatterer - 0.8.0.86
DMagic Orbital Science - 0.9.1
Editor Extensions - 2.5
Ferram Aerospace Research - 0.14.6
RasterPropMonitor - 0.18.3
KSP-AVC Plugin - 1.1.5
Kerbal Joint Reinforcement - 2.4.5
PlanetShine - 0.2.2.1
Procedural Parts - 0.9.21
RCS Build Aid - 0.5.4
RealChute - 1.2.6.3
RealSolarSystem - 8.5
RealismOverhaul - 7.0.4
RemoteTech - 1.5.2
SCANsat - 1.0.8.1
TextureReplacer - 2.1.2
TAC Life Support - 0.10.1.1320
Transfer Window Planner - 1.2.2

Screenshot of the rocket featured in my logs (in version 2.4.5):

Repro steps:

Install KJR 3 or 3.0.1

Load my save

Go to this rocket on the pad

Watch everything explode

Reinstall KJR 2.4.5

Load save

Go to this rocket on the pad

Nothing happens

Edited December 30, 2014 by ErrHead

[Urgante]

Same bug than ErrHead, windows 32 bit opengl, borderless window.

Big ships disassemble on load, smaller ones seem to work but they are constantly under acceleration.

Restoring 2.4.5 and all work perfectly.

Edited December 30, 2014 by Urgante

[WuphonsReach]

The primary decouplers that I use are: TR-2V (0.625m part), TR-18A (1.25m part) and Rockomax Brand Decoupler. No use of the TR-38-D part.

I was using 2.4.5 previously and upgraded to Kerbal_Joint_Reinforcement-v3.0.zip. Going to downgrade to 2.4.5 again.

[Damaske]

Ferram4, I am having troubles with the new update and sounding rockets not letting go of each other. I do have a lot of mods installed and Have yet tried to trace the trouble down yet! I plan to do that today or soonÂÂâ„¢. But other then the KSP.log file and a mod list what else is needed for the bug report?