Colliders are a bit bugged with this update?

[KerikBalm]

I've used some of my old designs that were tight/cramped fits but worked well in 1.3, and now I've often having bits go flying off in the weirdest ways.

I wonder if its also related to the issues some have had with docking (where a bump that doesn't result in docking can fling the 2 craft apart at a velocity many times greater than the bump).

Take this craft for instance:

Its a rather tight fit

But for some reason, the right (left on this image) gigantor near the docking port oftenbreaks off as it passes through the ramp-> cargobay connection.

It shouldn't be hitting anything, if the ISRU doesn't bump anything, then that gigantor shouldn't. It happens even when the ramp is at a shallow angle (its not like its hitting the roof). 

It doesn't make any sense why its breaking off, but I'm finding many of my cramped 1.3 designs have their breakable bits breaking off in 1.4.

Has anyone else noticed this?

[AVaughan]

Someone else posted a comment about deployable solar panel break when using the cargo ramp a few days ago.  (Unfortunately I can't find that post atm).  Hopefully this will be fixed for 1.4.3

[KerikBalm]

Has it been reported as a bug... I'm not even sure how to report this.

Here's an image sequence that shows it better:

(again, this used to work just fine)

Looks like it should fit, its a tight fit for the rear gigantors, but it should fit)

But nope, the one on the rear right (assuming the visible docking port is the front of the rover) of the rover breaks:

Well, that ones was a bit close to the edge, maybe this is forgivable, no?

The ones near the front surely shouldn't break right, they aren't sticking out like the rear one was:

What the heck?

That ain't right!

And why are they always breaking on just that one side???

[KerikBalm]

Yea... I'm not sure why that ended up on this thread... but no one else is finding extenable bits breaking when they shouldn't?

[KerikBalm]

Ok, so I wanted to see if this was fixed in 1.4.3, and its not, and I think its specific to the cargo ramp.

Take this rover:

Backing out too fast breaks off one of the gigantors (the one on the left of this image) but not the other. Going back in too fast... destroys the Science Jr??? what? why?

Why would the gigantors and science Jr break?

What is causing these collisions?! are they colliding with the node where the cargo-ramp attaches? What could they have changed in 1.4 to cause things in the empty space in the middle of the cargo-ramp to register as colliding with the cargo-ramp?

 

*edit* just tested with a long tunnel of mk3 cargobays... its only the cargo ramp causing this

Edited April 28, 2018 by KerikBalm

[klgraham1013]

Well.  Time to wait for 1.4.4.

[ExtremeSquared]

Have had same problem with small deployable solar panels and regular fairings. Non-outermost part, large clearance, still breaks. Larger clearance/clamshell/larger ejection force seemed to help, but it is still unexpected behaviour.

Edited July 20, 2018 by ExtremeSquared

[klgraham1013]

This was confirmed to be a feature and most certainly not a bug.

 

[KerikBalm]

Well.... I'd be interested in hearing that explanation that has apparently been provided "several times"

Quote

It's not a bug. But what was done in

... this certainly seems to me like its a bug.

You don't implement a partial solution unless you acknowledge that it was a problem (ie, a bug)

[klgraham1013]

4 minutes ago, KerikBalm said:

Well.... I'd be interested in hearing that explanation that has apparently been provided "several times"

Me too.  I actually don't recall seeing these explanations. 

Edited October 17, 2018 by klgraham1013

[XLjedi]

I need to check if it's any better in the 1.5 release.  Hopefully it is...

I did recently record a mission where I was able to disembark from a Mk3 cargo bay without destroying the solar panels.  But I recorded that pre 1.5?  ...so I dunno, maybe they snuck in a fix for it somewhere?

The bugtracker item that I submitted a while ago regarding this issue is here:

https://bugs.kerbalspaceprogram.com/issues/18017

I was assured (borderline ridiculed) by some that the coliders were all fine and that surely it was craft design and/or user error on my exit speed.  Even though the only thing about my craft design, the Mk3 Cargo Bay, or my exit speed that got changed in 1.4...  was the Unity engine. 

 

 

Edited October 19, 2018 by XLjedi

[KerikBalm]

Well, exit speed is an issue/workaround, I found that if I went through slow enough, it was ok. The limit was a couple m/s though. That's a problem on some heavier rovers that relied on having a bit of momentum to get up the ramp (sure I could make things lighter, but then part count goes up, and I shouldn't have to slowly creep something  high part count object up the ramp because of bad colliders)

[klgraham1013]

Adjusting speed is still just ignoring the issue though.  Which Squad seems very much content on doing.  It doesn't matter how fast I drive through a tunnel if there aren't invisible walls trying to tear off my side mirrors.

In stock 1.5.0.

Spoiler

 

 

Edited October 19, 2018 by klgraham1013

[XLjedi]

@klgraham1013 

I'll say it first...

That's by design and functioning as intended.

[MechBFP]

@JPLRepo Can you link us to those explinations you mentioned? I can’t find anything either. 

[XLjedi]

@MechBFP

If I'm not mistaken...  may have been in reference to my previous rant on the topic? 

In which @JPLRepo conclusion was:  I don't know how to back out of a Mk3 cargo bay and it's all good!

Unless there was some other more formal discussion on the "feature" that I have since missed.  I kinda gave up on it, so haven't tested it again lately...  I assume from @klgraham1013 video above that it's still works as intended.  Assuming the intention is to apply a filter for deployable parts on rovers. 

Edited October 19, 2018 by XLjedi

[klgraham1013]

All my tests confirm it's still "working as intended."  It does seem slightly better, but recreating the "Not a bug" is still stupidly easy.

Edited October 19, 2018 by klgraham1013

[KerikBalm]

Has anyone else replicated the asymmetry of this effect as well? It resulted in me making some unnecessarily asymmetrical designs, just because breakable bits break off on the right side but not the left:

This modular base had stuff on the top and right break off, but not low center/right break off.... so you see I've got the extendable parts either on the left (as they'd go in to dock) or low (less than 0.625m from the center). Those segments I could roll through the bay basically just fine... others, not so much.

Visually, and for non-extendable parts, you've got a 2.5m or more to work with (depending on height vs width), but not with the extendable parts, those you've got to fit them in a 1.25m diameter circle in the center, or to the left... because... that's the way they intended it to be... surely.

I guess I'll just increase the part collision tolerance if I can't creep the module through at 0.1 m/s

[klgraham1013]

I've done enough testing for @SQUAD on the matter.  More so then they seem to have.  They've got videos and persistent files.  They say the behavior is not a bug.  At this point, I'm done with it, and almost done with this whole thing.  I'm back to working on my own 1.3.2, which I should have never stopped working on apparently.  I'm getting really tired after 5+ years of this game constantly feeling unfinished.

[KerikBalm]

I saw your other bug reports with the deployment of a craft from a mk3 plane while landed, and I'm a bit worried it will affect some of my designs that i haven't tested since 1.3...

Specifically, my submarine deployer:

Spoiler

Honestly, I've barely played it on 1.4 (and not at all 1.5), due to being on an Arma 3 kick at the moment...

1.3 seemed complete to me (mods could supply what I was missing). I haven't seen anything in the last 2 updates that really interests me, and if its breaking/degrading the ability to deploy surface craft, then I think I'll just go back to 1.3.

Sure its a spaceflight game, but its doing things on the surface that kept my interest, otherwise I can just sit around in orbit forever. The challenge of getting payloadds to orbit, on a transfer orbit, and down to the surface was fun. Now, they're degrading the last step... bringing things back to the days when you'd just land a craft that is stationary on the surface, get out, plant a flag, and leave.

Well, there's no challenge in that for me anymore.

*edit"

Also, it seems clear to me from the video what the problem was... that your craft must have been placed on rails outside of the landed state (gets deleted). All craft on rails get deleted if they go in atmosphere that is too thick (corresponds to about 23km on Kerbin IIRC)

Edited October 30, 2018 by KerikBalm

[XLjedi]

6 minutes ago, KerikBalm said:

Sure its a spaceflight game, but its doing things on the surface that kept my interest, otherwise I can just sit around in orbit forever. The challenge of getting payloadds to orbit, on a transfer orbit, and down to the surface was fun. Now, they're degrading the last step... bringing things back to the days when you'd just land a craft that is stationary on the surface, get out, plant a flag, and leave.

Your playstyle sounds very much like my own.  I spend more time tinkering with craft on the surface that can dock/deploy and work as a team to accomplish missions, than I do launching singular rockets to a moon and back.  In my case, I also tend to favor spaceplanes over rockets.

Edited October 30, 2018 by XLjedi

[Dafni]

1 hour ago, klgraham1013 said:

...working on my own 1.3.2...

Color me intrigued!

A community based continuation of 1.3.1 crossed my mind more than once. But alas my coding skills are inexistent. In the meantime I consider a decently modded 1.3.1 install just that: a community supported "final" version of KSP, 1.3.2 if you will.

[klgraham1013]

10 hours ago, Dafni said:

Color me intrigued!

I'll try not to derail the thread to much here.

My main goal was to replace Squad's visuals and do a much needed, but likely never to happen cost balance pass.  The scope of what I intended has grown somewhat.

What I'm doing couldn't really be a mod, as I'm completely deleting Squad parts and replacing them with mod alternatives.  For instance, adding Ven's parts without module manager.  Basically doing what MM would do to the config, and then removing Squads models and textures.  I'm also removing redundant parts and parts with no real place in the whole.  I am then adding mod parts to fill out certain areas.  There are a few parts that feel like they were created without purpose, or consideration for how they interact with the rest, and only because the developers fancied the idea at the time.  I also removed non-lego parts, such as the Mk3 wings and Goliath.  There are mod alternatives that actually stick to the lego ethos of KSP and aren't model kit parts.  The ram usage is virtually unchanged.

I then worked on renaming parts so they were actually grouped properly; together with similar parts and in a logical order.  Consider the organization of the fuel tanks.  Shouldn't they be grouped in diameter, and then height?  Not the case in stock due to Squad's naming conventions.

The problem of cost balance is face palm worthy once you start looking into it.  A ladder that costs more than a probe core, solar panel, or rocket engine.  The crew cabin cost ratios are all over the place.  Especially when compared to command pods.  There's a lot of work to be done, but I'm starting to see the finish line.

I'm well aware this isn't really work I can share, since I'm not using Module Manager.  Really, I just want a game that I don't have to question.  That I don't have to wonder why in the world the Mk3 crew cabin is the 2nd most expensive part in the game (30,000 funds).  Only surpassed by the Mammoth (39,000).  More expensive than an Ion engine (8,000), an RTG (23,300), and the largest Xenon container (24,300).

[OHara]

17 hours ago, KerikBalm said:

Has anyone else replicated the asymmetry of this effect as well?

Yes.  Always the starboard side.  That fact was mentioned in the bug-report that XLJedi posted (18017) and I just noted that fact while confirming the one you posted (18801).

The earlier bug is marked "being worked on", and I am more willing to believe that than to believe hot-headed comments in the announcements thread.  Also, more speed seems to be required to cause breakage in version 1.5.1, about 5m/s now. On the other hand, this bug is weird enough and appeared when Squad wan't making any obvious changes to the affected parts, that I suspect the problem is deep in the Unity engine and might be beyond Squad's current capabilities to fix or avoid.

I'm with you in version 1.3.1  90% of the fun with 10% the frustration. 

(On the side topic of the submarine bug,  that was as you surmised, confusion about on-rails craft in the atmosphere mistakenly not being seen as 'splashed' and therefore deleted as if on re-entry, and was fixed in version 1.4.5, so fear not for your sub.)

[KerikBalm]

Ah, and you must be the very same person that confirmed the bug in the bug tracker:

Quote

Squad Bugtracker
https://bugs.kerbalspaceprogram.com

Issue #18801 has been updated by [email protected].

• Status changed from New to Confirmed
• % Done changed from 0 to 10

Confirming that constructions similar to those in the original-report's images, in version 1.5.1 still have retracted panels break when the seem they should be clear of the cargo bay walls. See related bug #18017 for example craft. The speeds required for breakage do seem to be higher, now about 5m/s, double what they were in 1.4.3.

The damage seems to occur at the transition between the MK3 cargo ramp and the hollow MK3 it is connected to, and always on the starboard side of the ramp (the side on starboard of an aircraft using this ramp as its tail).

A different example with maybe the same cause is described at https://forum.kerbalspaceprogram.com/index.php?/topic/177220-tips-for-making-stock-electricjet-powered-propellers-for-vtol-aircraft/
A helicopter craft also showed only the starboard-side elevons breaking, and being F3 reported as colliding with the MK3 ramp below.

My original report:

Quote

---

Bug #18801: Incorrect Collisions with mk3 Cargo Ramp

• Author: KerikBalm
• Status: Confirmed
• Priority: Low
• Assignee:
• Category: Physics
• Target version:
• Version: 1.4.3
• Platform: Windows
• Expansion: Core Game, Making History
• Language: English (US)

Parts will often break due to a collision when passing through the open space of the part, despite other parts passing through without collisions.

Some examples posted to Imgur:
A stock rover:
https://i.imgur.com/5sAe9as.png

The Science Jr. part colliding with the cargo ramp
https://i.imgur.com/81gnsnx.png

https://i.imgur.com/SwdXv0l.png

Redocking the rover, with the science Jr. now gone:
https://i.imgur.com/ZGuEqbt.png

A stock rover, with retracted "Gigantor" solar panels:
https://i.imgur.com/l7zkP9r.png

they break off when the "gigantor" intersects the plane of the rear of the cargobay:
https://i.imgur.com/K4t56R5.png

Note: after repeating this multiple times, it is always the one on the left side (from the point of view of this image).
another example, this time breaking off when loading instead of unloading a rover:
https://i.imgur.com/kLiRko4.png

https://i.imgur.com/4zfSna1.png
Note that it is again only 1 panel on the same side.

Passing through slowly enough avoids these effects