BUST: Kerbals as Debris

[Claw]

Bugs Under Special Test (BUST)

Kerbals as Debris

This thread is part of a series of community efforts to find the root cause of some of the more difficult to pinpoint bugs in KSP. The purpose of this thread is to for users to collaborate on solving the root of the problem.

Please see here for more details before posting.

Special Note: Please do not post requests for support on this thread. If you need help fixing a problem, please see the other Support stickies linked below.

Known Issues and Self Help

Support Request and Bug Reporting Guide

---

Purpose of the BUST threads:

• The BUST threads exist for discussion of the specific topic and for organization of the community of players who are collaborating to dig into the issue. The threads do not exist to answer support requests and any posts for support requests will likely be moved out of the thread.
  
• If you are familiar with the Bug Reporting Guidelines, what we are trying to do is solve the "How to replicate" portion of the bug report.

The ultimate goal of a BUST thread is to end up with a solid bug report that has easily repeatable steps. So if you post information, try to post clear directions on how you came up with it.

---

Topic:

Kerbals turning into debris.

Description:

This happens most often when crashing a rover with a kerbal attached to a command seat. The kerbal is ragdolled and dislodged from the seat and ends up labeled as "XXX debris".

There are at least two outcomes of the sort of crashing that induces this:

• Crashing results in a kerbal separate from any other component, but labeled as debris.
  
• Crashing results in an unresponsive kerbal still connected to the command seat, and unable to "leave" the seat
  

These might be related, or might not.

Primary Known Issue Threads:

http://forum.kerbalspaceprogram.com/threads/75586-Master-Thread-Unresponsive-Kerbals-in-EVA

What's known:

Kerbal's heads seem to be quite hard. Striking the helmet on the ground seems to be a contributor to turning the kerbal into debris. They have to strike their head hard enough that they are ragdolled and dislodged from the seat, but not so hard that they are killed and disappear from existence. It's unclear if the helmet needs to strike the ground first, or if it can strike after any other portion of the craft first.

A possible related topic is when a kerbal goes EVA and the game crashes when they touch the ground. This may also be related to ragdoll states.

Edited November 16, 2014 by Claw

[Claw]

So it's getting late but I thought the first attempts at this couldconsist of a small, rover like craft suspended upside down by a launch clamp. Send a kerbal EVA and put him into the command chair. After the crash test dummy is firmly seated, drop him (upside down, head first) into the pavement. Having it on a launch clamp makes the height adjustable to control how hard he hits the ground.

Hopefully someone can figure this out.

Cheers,

~Claw

[Wanderfound]

Think this is a separate issue from the post-ragdoll scarecrow bug, or all part of the same thing?

That was induced by just walking around on the runway; first one kerbal, then the other. In each case, they went into ragdoll after falling off the plane, then stayed stuck in scarecrow mode once they got up. Both Scott Manley and Kurtjmac have featured similar events in the videos.

[BagelRabbit]

I generally have it happen on my small rovers. Here's what I've found about this issue:

It has happened without the Kerbal's head touching the ground. If a Kerbal is ragdolled by crashing into another object (not even necessarily headfirst!), this sort of thing can happen. You just can't see it happen often, because it takes more specialized circumstances.

Here is a video of an unrelated bug (and one that still isn't easily explainable... if you have some time to head over here and check it out, it would be much appreciated.) [sorry, wasn't supposed to ask for help on this thread.] In this video, the Kerbal goes ragdoll after crashing. What isn't shown, but what happened on three separate occasions, was that the Kerbal remained 'ragdoll-ed' in the seat, and was counted as debris.

It seems to be related to clipping the Command Seat into other parts. It happens much of the time when I have the Kerbal clipped into another part, just a small ways.

It seems to be related to having the Kerbal stay "in" the seat even after it goes ragdoll. This often happens because of part clipping.

When the Kerbal crashes in the correct way to bring about the bug, it will not get up, but remain in Ragdoll mode until another part is switched to. When another part is switched to, the Kerbal generally gets up, but when he is switched to, he cannot do anything.

Would it help to provide a .craft file for a rover that always induces the effect? It's small (13 parts) and so it would be reasonably easy to see which bit was causing this ragdoll effect.

Think this is a separate issue from the post-ragdoll scarecrow bug, or all part of the same thing?

I'm almost positive it's separate, as the scarecrow bug is related in no way to External Command Seats, which is the root of the Kerbals-as-Debris bug. It certainly is something that I would like to have fixed though, maybe Claw will make a BUST thread for it after investigations of this bug are concluded.

Edited November 16, 2014 by UpsilonAerospace

[Claw]

Think this is a separate issue from the post-ragdoll scarecrow bug, or all part of the same thing?

That was induced by just walking around on the runway; first one kerbal, then the other. In each case, they went into ragdoll after falling off the plane, then stayed stuck in scarecrow mode once they got up. Both Scott Manley and Kurtjmac have featured similar events in the videos.

I don't know. It's probably separate but maybe... That's why we're here! I thought frozen-T kerbals had gone away. Obviously you have a couple mods there. We'd have to narrow down if this is a stock problem, mod induced stock problem, or just a mod problem. It probably is a separate issue though because of how the kerbals behave. (Just confirm you're using v0.25?)

It has happened without the Kerbal's head touching the ground. If a Kerbal is ragdolled by crashing into another object (not even necessarily headfirst!), this sort of thing can happen. You just can't see it happen often, because it takes more specialized circumstances.

Outlinking for help is fine (and probably encouraged). I just didn't want these threads to turn into support/help threads (because we have a whole forum for that!)

Here is a video of an unrelated bug (and one that still isn't easily explainable... if you have some time to head over here and check it out, it would be much appreciated.) [sorry, wasn't supposed to ask for help on this thread.]

Outlinking for help is fine (and probably encouraged). I just didn't want these threads to turn into support/help threads (because we have a whole forum for that!)

In this video, the Kerbal goes ragdoll after crashing. What isn't shown, but what happened on three separate occasions, was that the Kerbal remained 'ragdoll-ed' in the seat, and was counted as debris.

Yes. There seems to be two things that can happen here. Kerbals ragdolled while still connected to the command seat, and kerbals that get ejected and turned into debris. I forgot all about mentioning the kerbal-in-the-seat problem. Updating the OP.

Would it help to provide a .craft file for a rover that always induces the effect? It's small (13 parts) and so it would be reasonably easy to see which bit was causing this ragdoll effect.

Most definitely. If you have a way to make it happen nearly 100% of the time, then great!

Cheers,

~Claw

[BagelRabbit]

All right. Here's the craft file. The rover is called Pico-Pico-Pico Rove B on account of its small size.

How to induce the effect:

Decouple the rover and drive around. The rover has a high CoM and thus is relatively tippable, especially off the side of the runway. If you tip the rover at a high enough speed, the Kerbal will ragdoll but still more-or-less remain in the seat. Switching to another object and back makes the Kerbal get up (it may take several attempts), but when he's up, he's counted as debris. Switching to the Kerbal, he has his own navball and everything! Also, you can assess his Monopropellant levels but can't take surface samples, pull out the EVA pack, or move in general.

Here's a picture of Bob staring out into space near the crashed rover. Hello, Bob? Earth to Bob! oh, not again.

I have a feeling that this will make a good testbed for Debris-Kerbals. See if you can't find a way to fix the problem.

-Upsilon

[Claw]

Cool! I haven't given your rover a try, but here is my progress so far.

I haven't done much testing yet to see how sensitive it is to height, but out of 5 drops so far, this turned the kerbal into debris 5 times.

Going EVA and releasing from the ladder did ragdoll the kerbal, but did not kill or break him.

However, the seat is a different story:

1. So you can see how the contraption is built.

2. Send the kerbal on EVA and board the seat (can be done from the pod ladder or from the ground).

3. Stage the craft and let it crash to he ground (head first. Sorry Jeb!)

The kerbal is now debris.

Javascript is disabled. View full album

Edited November 16, 2014 by Claw

[Wanderfound]

I don't know. It's probably separate but maybe... That's why we're here! I thought frozen-T kerbals had gone away. Obviously you have a couple mods there. We'd have to narrow down if this is a stock problem, mod induced stock problem, or just a mod problem. It probably is a separate issue though because of how the kerbals behave. (Just confirm you're using v0.25?)

That happened around the end of August, so it would've been .24 at the time. SP+, TAC-LS, TAC-FB, FAR, Kerbpaint, Mechjeb and the usual basic utility mods (RCS Build Aid, EE, PAD, KAC, etc). Can't get more precise than that without a time machine, unfortunately.

I don't think I've had it happen more than a couple of times, and not recently.

Edited November 16, 2014 by Wanderfound

[xEvilReeperx]

I haven't done much testing yet to see how sensitive it is to height, but out of 5 drops so far, this turned the kerbal into debris 5 times.

I haven't tested anything, but my money would be that whatever logic happens when the kerbal gets out of the seat isn't happening when the joint between kerbal and seat is broken. You end up with presumably a separate (but not commandable; no ModuleCommand) Vessel that ends up auto-classified as debris.

It might be as simple as adding a PartModule that's aware of the command seat status of the Kerbal and calling something like KerbalEVA.OnDeboardSeat() when MonoBehaviour.OnJointBreak occurs

[Claw]

I haven't tested anything, but my money would be that whatever logic happens when the kerbal gets out of the seat isn't happening when the joint between kerbal and seat is broken. You end up with presumably a separate (but not commandable; no ModuleCommand) Vessel that ends up auto-classified as debris.

Yeah, that's probably exactly what's happening. The kerbal looks like any other part when attached to the command seat. So he's simply breaking off and turning into debris. Although I think there might be more to it, because sometimes they are classified as a "base."

Part of the problem with slapping a fix in though is that the name is already changed and a few other things occur. Might be able to look him up in the roster, but there's some knowledge that needs to go into relabeling him as a kerbal (vice debris).

~Claw

[xEvilReeperx]

The part itself should have the info you need (Part.protoModuleCrew)

[Claw]

There is now a fix posted here: http://forum.kerbalspaceprogram.com/threads/97285-0-25-Stock-Bug-Fix-Modules-%28Release-v0-1-5-22-Nov-14%29

This will prevent kerbals from turning into debris. If you already have kerbals that are debris, this will recover them.

Cheers,

~Claw

Edited November 23, 2014 by Claw