Fixed Decoupler for KSP x64 on Windows - v4 - (July 22)

[Laminator]

A quick update about my testing: PF does not work, but good old KW Fairings work with this fix. At least in my full blown install (~60 mods installed)

[Kolago]

Yes, PF 3.06 does not work in the moment.

[ragzilla]

Hey sarbian, can you add a notice in the OP that you have to remove/readd decouplers to existing craft to get this to apply? Might also want to mention that craft already in flight won't work either.

[OrbitalDebris]

Awesome stuff! Thanks Sarbian!

[sharpspoonful]

I think that mods like PF, TweakableEverythink and KAS must implement a fix like this into there mods.

Does this fix the unity decouple() bug? Will this fix the errors with KAS? If so you might want to contact the mod maker of Kerbal Attachment System.

No, this just fixes the lack of force from decouplers, not the (decouple) node. That is a separate issue that was brought about by the way x64 handles the code. Majir has brought this issue to light several times, as seen in the x64 compatibility and the KAS threads. That fix is going to have to be applied more or less likely by SQUAD or the Unity team, and is not something that the modders themselves will be able to fix.

Edited July 21, 2014 by sharpspoonful

[Kolago]

So the force to jettison the fairings is not related to the fix you made?

[sidfu]

So the force to jettison the fairings is not related to the fix you made?

probaly for fairings u gonna have to go to the .cfg for it and add the new module

[sharpspoonful]

So the force to jettison the fairings is not related to the fix you made?

I didn't make it, but like I said, this is an issue that has been addressed by Majir in several thread. PF 3.06 probably has an issue with how its applying the the decouple mode. I have to say, I have no idea.

[DasBananenbrot]

Wait, so stock decouplers have a bug but shouldn't that also affect mod decouplers then? That would be a bummer because getting for fixes for all the mods would take some time

[biohazard15]

Sarbian, thanks for this!

[sidfu]

was looking at differn fairings with the setup of this mod it should replace the decoupler module for kw farings since it uses the stock decoupler. procedral fairings thou use a custom decoupler so probaly wotn work on them. seems like novapunch also uses stock decoupler module. havaent tried it yet thou.

[biohazard15]

was looking at differn fairings with the setup of this mod it should replace the decoupler module for kw farings since it uses the stock decoupler. procedral fairings thou use a custom decoupler so probaly wotn work on them. seems like novapunch also uses stock decoupler module. havaent tried it yet thou.

Any part that uses stock decoupler module will be fixed by this mod.

Sarbian, can you shed some light on this bug? Is it on Unity or Squad side? How hard exactly it was to fix it? Can Squad do the same thing in future updates?

[sarbian]

About >MY< replacement decoupler not working with the KerbalX : it's a bug with .24 and ship older than .24

.24 had a change in the module loading to fix a problem when the module are not in the same order in a part and in a saved ship. It seems that fix make things worse when the module count is not the same (aka old save since most part in .24 had module added to track contract).

So when you load a KerbalX the module are messed up. You can see ii in your logs with those messages : Attempting to replace PartModule 'XXX' with 'YYY'

biohazard15 : the bug (as far as I can tell) is that you can to a decouple and apply force to a part in the same physic frame (at the same time if it makes more sens to you). My decoupler wait 1 frame (about 0.04 seconds) to apply the forces.

Edited July 20, 2014 by sarbian

[Gaalidas]

That was fast too, sorta makes you wonder why it wasn't fixed already on the official side of things. Don't they test this stuff?

[sidfu]

looks like tweakeverything messes this up and maybe kerbpaint also causes it not to work right so might want to add to op.

[toadicus]

All: the issue with TweakableEverything should be fixed by removing the ":FOR[DecoupleForX64]" from the @PART lines in DecoupleForX64.cfg and changing the decouplerModuleName for stack decouplers. ModuleManager doesn't allow multiple pass declarations and just uses the first; by saying ":FOR[DecoupleForX64]:AFTER[TweakableEverything]", the patch applies during the ":FOR[DecoupleForX64]" pass, long before the TweakableEverything pass. Also, the DecoupleForX64 cfg mistakenly used ModuleAnchoredDecouplerX64 for the ModuleDecoupleX64 decouplerModuleName; just a copypasta mistake.

TE and DFX64 should play nicely with this cfg:

@PART[*]:HAS[@MODULE[ModuleDecouple]]:AFTER[TweakableEverything]
{
@MODULE[ModuleDecouple]
{
@name = ModuleDecoupleX64
}

@MODULE[ModuleTweakableDecouple]
{
@decouplerModuleName = ModuleDecoupleX64
}
}

@PART[*]:HAS[@MODULE[ModuleAnchoredDecoupler]]:AFTER[TweakableEverything]
{
@MODULE[ModuleAnchoredDecoupler]
{
@name = ModuleAnchoredDecouplerX64
}

@MODULE[ModuleTweakableDecouple]
{
@decouplerModuleName = ModuleAnchoredDecouplerX64
}
}

[FenrirWolf]

Cool. I'll have to give this a shot next time I play.

[Woodstar]

I never noticed this bug. Guess I'll check it out tonight, I'm running 64bit exe also.

[Kamuchi]

All: the issue with TweakableEverything should be fixed by removing the ":FOR[DecoupleForX64]" from the @PART lines in DecoupleForX64.cfg and changing the decouplerModuleName for stack decouplers. ModuleManager doesn't allow multiple pass declarations and just uses the first; by saying ":FOR[DecoupleForX64]:AFTER[TweakableEverything]", the patch applies during the ":FOR[DecoupleForX64]" pass, long before the TweakableEverything pass. Also, the DecoupleForX64 cfg mistakenly used ModuleAnchoredDecouplerX64 for the ModuleDecoupleX64 decouplerModuleName; just a copypasta mistake.

TE and DFX64 should play nicely with this cfg:

@PART
[*]:HAS[@MODULE[ModuleDecouple]]:AFTER[TweakableEverything]
{
@MODULE[ModuleDecouple]
{
@name = ModuleDecoupleX64
}

@MODULE[ModuleTweakableDecouple]
{
@decouplerModuleName = ModuleDecoupleX64
}
}

@PART
[*]:HAS[@MODULE[ModuleAnchoredDecoupler]]:AFTER[TweakableEverything]
{
@MODULE[ModuleAnchoredDecoupler]
{
@name = ModuleAnchoredDecouplerX64
}

@MODULE[ModuleTweakableDecouple]
{
@decouplerModuleName = ModuleAnchoredDecouplerX64
}
}

Thank you!

[sarbian]

All: the issue with TweakableEverything should be fixed by removing the ":FOR[DecoupleForX64]" from the @PART lines in DecoupleForX64.cfg and changing the decouplerModuleName for stack decouplers. ModuleManager doesn't allow multiple pass declarations and just uses the first; by saying ":FOR[DecoupleForX64]:AFTER[TweakableEverything]", the patch applies during the ":FOR[DecoupleForX64]" pass, long before the TweakableEverything pass. Also, the DecoupleForX64 cfg mistakenly used ModuleAnchoredDecouplerX64 for the ModuleDecoupleX64 decouplerModuleName; just a copypasta mistake.

Thanks for writing better MM patch than me toadicus

I'll update the zip when I can.

[NoJeb]

Great idea

Problem... it Crashed KSP in the VAB while trying to load a Ship (it's a ship i made before the mod. it's nothing fancy only stock parts....

Now the savegame is wrecked and will crash KSP... sight.. gone are my 10 mil funds ^^

[KerbMav]

So, would this work?

no it wont, deleted before anything...

Edited July 21, 2014 by KerbMav

[koksny]

So, would this work?

No, it does not. We need to fork Fixed Decoupler into this:

[KSPEvent(name = "Jettison", active=true, guiActive=true, guiActiveUnfocused=false, guiName="Jettison")]

public void Jettison()

{

if (part.parent)

{

foreach (var p in part.parent.children)

foreach (var joint in p.GetComponents<ConfigurableJoint>())

if (joint!=null && (joint.rigidbody==part.Rigidbody || joint.connectedBody==part.Rigidbody))

Destroy(joint);

part.decouple(0);

var tr=part.FindModelTransform(transformName);

if (tr)

{

part.Rigidbody.AddForce(tr.TransformDirection(forceVector)

*Mathf.Lerp(ejectionLowDv, ejectionDv, ejectionPower),

ForceMode.VelocityChange);

part.Rigidbody.AddTorque(tr.TransformDirection(torqueVector)

*Mathf.Lerp(ejectionLowTorque, ejectionTorque, torqueAmount),

ForceMode.VelocityChange);

}

else

Debug.LogError("[ProceduralFairingDecoupler] no '"+transformName+"' transform in part", part);

ejectFx.audio.Play();

}

}

I have tried adding a loop with frame count, but it didn't work. Any help from some actual coder?

EDIT: Also atm using Fixed Decoupler messes with launch stability enhancers, might be good idea to make exception to them.

Edited July 21, 2014 by koksny

[Sesambrot]

No, it does not. We need to fork Fixed Decoupler into this:

I have tried adding a loop with frame count, but it didn't work. Any help from some actual coder?

EDIT: Also atm using Fixed Decoupler messes with launch stability enhancers, might be good idea to make exception to them.

You know, I'm no coder either, but just to be safe anyone wanting to take a shot at this, here's the whole thing:

https://github.com/e-dog/ProceduralFairings/blob/master/Source/FairingDecoupler.cs

And just to add some context; the original creator is refusing to fix it because Squad didn't implement procedural cost generation...

Make of that what you will, but for someone who knows what he's doing this is probably rather simple.

And just so everybody is aware of what exactly you're dealing with here:

The mod in question: http://forum.kerbalspaceprogram.com/threads/39512-0-24-Procedural-Fairings-3-06-%28July-20%29

It's liscence: http://creativecommons.org/licenses/by/3.0/

[Borklund]

Thanks sarbian. I really don't understand what Squad are doing in the 100-200 day periods between each update.