Compund_Collision_chain demo not working properly #85932

Closed
opened 2021-02-23 23:01:27 +01:00 by Nahuel Belich · 21 comments

System Information
Operating system: Linux-5.4.0-65-generic-x86_64-with-debian-bullseye-sid 64 Bits
Graphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.32.03

Blender Version
Broken: version: 2.92.0 Release Candidate, branch: master, commit date: 2021-02-22 12:21, hash: e497c1b93c
Worked: (newest version of Blender that worked as expected)

Short description of error
The demo file Compund_Collision_chain that can be downloaded from https://www.blender.org/download/demo-files/ it doesn't seem to be working properly

Exact steps for others to reproduce the error
1- Download the file compound_collision_chain.blend from https://www.blender.org/download/demo-files/

2-in the properties tab "scene properties" delete the bake for the rigid bodies image.png

3- Probably after deleting the bake the baked keyframes are still in ram if you did playback before deleting the cache, add a keyframe to the position of the first chain link just to force blender to dump the cache.

4- This should happend if you bake again or try to playback the animation puff.mp4

**System Information** Operating system: Linux-5.4.0-65-generic-x86_64-with-debian-bullseye-sid 64 Bits Graphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.32.03 **Blender Version** Broken: version: 2.92.0 Release Candidate, branch: master, commit date: 2021-02-22 12:21, hash: `e497c1b93c` Worked: (newest version of Blender that worked as expected) **Short description of error** The demo file Compund_Collision_chain that can be downloaded from https://www.blender.org/download/demo-files/ it doesn't seem to be working properly **Exact steps for others to reproduce the error** 1- Download the file compound_collision_chain.blend from https://www.blender.org/download/demo-files/ 2-in the properties tab "scene properties" delete the bake for the rigid bodies ![image.png](https://archive.blender.org/developer/F9832711/image.png) 3- Probably after deleting the bake the baked keyframes are still in ram if you did playback before deleting the cache, add a keyframe to the position of the first chain link just to force blender to dump the cache. 4- This should happend if you bake again or try to playback the animation [puff.mp4](https://archive.blender.org/developer/F9832723/puff.mp4)
Author

Added subscriber: @NahuelBelich

Added subscriber: @NahuelBelich
Member

Added subscribers: @ZedDB, @lichtwerk

Added subscribers: @ZedDB, @lichtwerk
Member

Changed status from 'Needs Triage' to: 'Confirmed'

Changed status from 'Needs Triage' to: 'Confirmed'
Member

Can confirm.

@ZedDB : does this ring a bell?

Can confirm. @ZedDB : does this ring a bell?

Seems like something is wrong with the hosted test file!

If you use the original file from the patch, it works: https://developer.blender.org/D5797

I've have seen this issue before and it seems like the bullet physics world has somehow gotten into an invalid state.
So no matter what you do, things will explode and get "ghost" forces applied to them.

Previously I had to delete the rigidbody physics world and start over in the file to fix this.
I don't know why or how this happens, perhaps it has to do with compression as the original example file is ~6mb while the newer broken one is ~800kb...

Seems like something is wrong with the hosted test file! If you use the original file from the patch, it works: https://developer.blender.org/D5797 I've have seen this issue before and it seems like the bullet physics world has somehow gotten into an invalid state. So no matter what you do, things will explode and get "ghost" forces applied to them. Previously I had to delete the rigidbody physics world and start over in the file to fix this. I don't know why or how this happens, perhaps it has to do with compression as the original example file is ~6mb while the newer broken one is ~800kb...

@NahuelBelich can you see if you can reproduce the issue by starting with the original working file and then try to make it break like in the new example file?
I'm guessing it is related to saving and loading the file.

@NahuelBelich can you see if you can reproduce the issue by starting with the original working file and then try to make it break like in the new example file? I'm guessing it is related to saving and loading the file.
Author

@ZedDB ill try tonight and let you know

@ZedDB ill try tonight and let you know

@NahuelBelich Great!
Note the I have never myself been able to reproduce broken files it is only from bug reports that I have gotten them.
So if you could figure out how to break it that would be super!

@NahuelBelich Great! Note the I have never myself been able to reproduce broken files it is only from bug reports that I have gotten them. So if you could figure out how to break it that would be super!
Author

@ZedDB
This is becoming a bit weirder than my artist brain can understand, first the tests.


Blend file "Comptest Chain Original.blend" From https://developer.blender.org/D5797

2.91.2 : Works fine
2.92 candidate : Works fine
2.93 alpha : Works fine


Blend File "compound_collision_chain.blend" from https://www.blender.org/download/demo-files/

2.91.2 : fails
2.92 Candidate : fails
2.93 alpha : fails

This is the weirdest part, i though that some property could be different between objets in the files and the properties looks all the same, then i though may be the compress option for the blend file, didn't change anything, then i try this.

In the old demo the one that weights 6mb
1 - select a few links of the chain
2- move it a bit to one side
3- Ctrl+c to copy
4- Ctrl+v to paste the object in a new empty blend file
did the same with the file hosted in blender.org the one that weights 769 Kb, move the chains same way but to the other side, shoud look the same under a new empty blender but. .
Chains demo.blend
asdasdad.mp4
as you can see the chains looks very different, the one on the right from the hosted blender.org file, twists into itself in a really weird way.
i check the objetcs one by one and the rigid bodies properties are the same at least in the interface that i can see, should look the same but it does not.

I made a test from scratch, I made new objects in a new file and copy all the parameters by hand for the rigid body from+ the chain in the hosted blender.org file, the ones that looks bad in the video above, and it works as a charm in this new file.
chain from scratch.blend
aaaaa.mp4

i wasn't able to crash the original file, save with compression with out compression, baked, no baked, not sure whats happening but as it looks what its breaking things can be copy with the objects if you copy objects from blend to blend.

@ZedDB This is becoming a bit weirder than my artist brain can understand, first the tests. ---- Blend file "Comptest Chain Original.blend" From https://developer.blender.org/D5797 2.91.2 : Works fine 2.92 candidate : Works fine 2.93 alpha : Works fine ---------------------- Blend File "compound_collision_chain.blend" from https://www.blender.org/download/demo-files/ 2.91.2 : fails 2.92 Candidate : fails 2.93 alpha : fails This is the weirdest part, i though that some property could be different between objets in the files and the properties looks all the same, then i though may be the compress option for the blend file, didn't change anything, then i try this. In the old demo the one that weights 6mb 1 - select a few links of the chain 2- move it a bit to one side 3- Ctrl+c to copy 4- Ctrl+v to paste the object in a new empty blend file did the same with the file hosted in blender.org the one that weights 769 Kb, move the chains same way but to the other side, shoud look the same under a new empty blender but. . [Chains demo.blend](https://archive.blender.org/developer/F9836879/Chains_demo.blend) [asdasdad.mp4](https://archive.blender.org/developer/F9836877/asdasdad.mp4) as you can see the chains looks very different, the one on the right from the hosted blender.org file, twists into itself in a really weird way. i check the objetcs one by one and the rigid bodies properties are the same at least in the interface that i can see, should look the same but it does not. I made a test from scratch, I made new objects in a new file and copy all the parameters by hand for the rigid body from+ the chain in the hosted blender.org file, the ones that looks bad in the video above, and it works as a charm in this new file. [chain from scratch.blend](https://archive.blender.org/developer/F9837129/chain_from_scratch.blend) [aaaaa.mp4](https://archive.blender.org/developer/F9837135/aaaaa.mp4) i wasn't able to crash the original file, save with compression with out compression, baked, no baked, not sure whats happening but as it looks what its breaking things can be copy with the objects if you copy objects from blend to blend.

Added subscriber: @Fernando-Alcala

Added subscriber: @Fernando-Alcala

David's original scene Comptest Chain.blend (6mb) used three separate cubes each with a convex hull. This one works.
comptest.png
Then Dalai did some "cleanup" compound_collision_chain.blend (777kb) and merged those cubes, creating a concave shape that still has convex hull. This is what you download off of Blender. Doesn't work.
blend comp coll.png
David then fixes this compound_collision_chain.blend (253kb) by separating the bottom cube and removing the rigid bodies from the wall.
comp collision David.png

David explains this in the last comment of that thread.

As for the chain behaving differently, I also couldn't find anything different. Same numbers, same parenting, same orientation, same structure, but obviously something happened. What you can see is that only two links of the top chain seem to be different. 2021-02-25 13-34-10.mp4

David's original scene *Comptest Chain.blend* (6mb) used three separate cubes each with a convex hull. This one works. ![comptest.png](https://archive.blender.org/developer/F9839670/comptest.png) Then Dalai did some "cleanup" *compound_collision_chain.blend* (777kb) and merged those cubes, creating a concave shape that still has convex hull. This is what you download off of Blender. Doesn't work. ![blend comp coll.png](https://archive.blender.org/developer/F9839673/blend_comp_coll.png) David then fixes this *compound_collision_chain.blend* (253kb) by separating the bottom cube and removing the rigid bodies from the wall. ![comp collision David.png](https://archive.blender.org/developer/F9839675/comp_collision_David.png) David explains this in the last comment of that thread. As for the chain behaving differently, I also couldn't find anything different. Same numbers, same parenting, same orientation, same structure, but obviously something happened. What you can see is that only two links of the top chain seem to be different. [2021-02-25 13-34-10.mp4](https://archive.blender.org/developer/F9839896/2021-02-25_13-34-10.mp4)

@Fernando-Alcala what thread are you talking about here. I'm a bit lost at it seems like you are quoting a forum thread but not linking to it.

@Fernando-Alcala what thread are you talking about here. I'm a bit lost at it seems like you are quoting a forum thread but not linking to it.

D5797#237065 comment section?

[D5797](https://archive.blender.org/developer/D5797)#237065 comment section?

Added subscribers: @sreich, @dfelinto

Added subscribers: @sreich, @dfelinto

Thanks for the report. Marking this as known issue since it depends on the rigid body system which at the moment has no active developer. I'm tagging Sergej here anyways in case he wants to tackle any of those. Anyone wlse interested on that can also join the geometry nodes meetings to get onboarded.

Thanks for the report. Marking this as known issue since it depends on the rigid body system which at the moment has no active developer. I'm tagging Sergej here anyways in case he wants to tackle any of those. Anyone wlse interested on that can also join the geometry nodes meetings to get onboarded.

@dfelinto This is not a know issue.
This is not a problem with the simulation code at all.
The issue here is that you did some cleanup of the test file that broke it.
To unbreak it, we should replace the current file we host with the original file that was provided to us.

@dfelinto This is not a know issue. This is not a problem with the simulation code at all. The issue here is that you did some cleanup of the test file that broke it. To unbreak it, we should replace the current file we host with the original file that was provided to us.
Member

Added subscriber: @HooglyBoogly

Added subscriber: @HooglyBoogly
Member

If it's not a known issue then I assume it can be classified as a bug? The idea was to complete the triaging process and be realistic about our ability to fix many of these reports.

If it's not a known issue then I assume it can be classified as a bug? The idea was to complete the triaging process and be realistic about our ability to fix many of these reports.

Changed status from 'Confirmed' to: 'Resolved'

Changed status from 'Confirmed' to: 'Resolved'
Dalai Felinto self-assigned this 2022-05-10 15:29:35 +02:00

Got it. I don't know what, but I just did just some minor cleanups and re-uploaded the file to the same place. If it doesn't work let me know. (this time I even uploaded the file un-baked)

Got it. I don't know what, but I just did just some minor cleanups and re-uploaded the file to the same place. If it doesn't work let me know. (this time I even uploaded the file un-baked)

Seems to be working now, thanks! :)

Seems to be working now, thanks! :)
Sign in to join this conversation.
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset Browser
Interest
Asset Browser Project Overview
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
Interest
EEVEE & Viewport
Interest
Freestyle
Interest
Geometry Nodes
Interest
Grease Pencil
Interest
ID Management
Interest
Images & Movies
Interest
Import Export
Interest
Line Art
Interest
Masking
Interest
Metal
Interest
Modeling
Interest
Modifiers
Interest
Motion Tracking
Interest
Nodes & Physics
Interest
OpenGL
Interest
Overlay
Interest
Overrides
Interest
Performance
Interest
Physics
Interest
Pipeline, Assets & IO
Interest
Platforms, Builds & Tests
Interest
Python API
Interest
Render & Cycles
Interest
Render Pipeline
Interest
Sculpt, Paint & Texture
Interest
Text Editor
Interest
Translations
Interest
Triaging
Interest
Undo
Interest
USD
Interest
User Interface
Interest
UV Editing
Interest
VFX & Video
Interest
Video Sequencer
Interest
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
Legacy
Blender 2.8 Project
Legacy
Milestone 1: Basic, Local Asset Browser
Legacy
OpenGL Error
Meta
Good First Issue
Meta
Papercut
Meta
Retrospective
Meta
Security
Module
Animation & Rigging
Module
Core
Module
Development Management
Module
EEVEE & Viewport
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline, Assets & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Priority
High
Priority
Low
Priority
Normal
Priority
Unbreak Now!
Status
Archived
Status
Confirmed
Status
Duplicate
Status
Needs Info from Developers
Status
Needs Information from User
Status
Needs Triage
Status
Resolved
Type
Bug
Type
Design
Type
Known Issue
Type
Patch
Type
Report
Type
To Do
No Milestone
No project
No Assignees
6 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#85932
No description provided.