Hello.
After updating my Android project from Unity 2019 to 2022 i have noticed that ragdolls is not working.
The raggdol is just standing in T pose and not falling. No scripts or animation attached.
What could be wrong?
Okay, he is falling in clean project so the problem is somewhere in settings.
And as i see all Rigidbodies is not working in the updated project.
Following this to see if you find out the underlying issue. We also got some similiar albeit different mysterious problems with Physics2D rigidbodies after updating to 2022.2 version. In a content that has been working fine initially from Unity version 2019 through to 2021.
In the end i removed the Project Settings folder from project and it is fixed the problem, but now i need manually set up all the settings.
paljan i hope you find a solution.
Experimentally i figured out that problem was in DynamicsManager.asset file. Removing it fixing the problem.
I don’t see why you would but would be interested in seeing a project that exhibits this. There were no such bugs reported for 2D physics. The ONLY thing that could cause that would be if there were no time-updates for FixedUpdate therefore no simulation. Box2D itself hasn’t changed at all in years which is the only way Rigidbody wouldn’t move.
Thanks OS13 and MelvMay! Well seeings as OS13 got his problem sorted out, maybe I dont mind hijacking his thread!
Our issue is a rather weird one: We have multiple projects with a collection of many minigames in them. Right after updating from Unity 2021.2 to 2022.2 (Now in 2022.2.3) different minigames in two of our projects started exhibiting similiar behaviour. The rigidbody2D objects seem to randomly get a huge impulse of force when being simulated freely and bouncing around. In one minigame you can drag and drop the physics objects and in the other you just spawn them with a touch.
Here is the peculiar part: this only happens in iOS builds. In the editor and in android builds everything still works like in the current store build which is built with 2021.
Now, reading OS13’s solution here, I also just now forced Unity to regenerate the DynamicsManager.asset file and only put the collision matrix back in place. This updated it from serializedVersion 13 → 14 and added and changed some things. Additionally this also affected the behaviour of the rigidbodies, although it didn’t fix them.
Now, after regenerating the DynamicsManager, in the minigame where you can drag and drop the rigidbodies, some of them dont get bounced around that hard anymore. When placed on the ground collider, they sort of twitch and with this twitching motion sometimes sink into the bottom collider and only from there get a big impulse of force. Earlier you could not observe this small scale twitching and sinking, rather only random smaller bounces and then a huge one always at a random collision. Big enough to blast them right through all the colliders and off screen.
We are trying to determine which minigame would be a better example to try and strip down for a possible bug report. As both of our applications are big projects with lot of content and minigames. I’ve attached the changes the regeneration triggered in the DynamicsManager as these somehow affected also the erroneous behaviour.
EDIT: Also, this guy seemed to have similiar issues, although he also had some funky logic. Our bounces happen also without any additional game logic affecting the rigidbodies, only simulated collisions: iOS build physics behave differently on the iphone
The “DynamicsManager” is the 3D physics settings, not the 2D physics settings and have nothing at all to do with 2D. There’s not a single thing in there that could affect 2D physics.
The impulse solver hasn’t changed in Box2D for years so I have no idea why you’re see random/sudden impulses. The only thing I could think of would be user forces are being applied in a frame-dependent manner and you’re occasionally seeing the effects of that. Beyond that, I have no idea.
Thank you for your thoughts MelvMay, although then I am even more flabbergasted. One of these minigames has indeed been first implemented and published in 2020 using Unity 2019 and has worked without any issues through multiple updates all the way up until now. We’ve started stripping down one of the projects to be able to isolate one of the problematic minigames and to be able to file a bug report, probably best to continue with that! Thank you for your time!
Edited: (see further reply below)
Getting strange behaviour on Windows (Unity 2022.2.4f), mostly with Physics . Seemed to happen after I imported a scene & some prefabs from a previous project (via drag-and-drop 2nd editor loaded, been doing that fine for years). Even creating fresh blank scenes have rigidbodies not working (at all) and I confirmed numerous times seeing default settings for Physics, time, etc…
At a guess, something is corrupting settings files? Creating unpredictable havok (at least Physics).
Deleting project settings (Edit: Scratch that, delete just the Physics related ones) seems to be the only thing that fixes it.
Interesting! I already tried regenerating all the physics settings related .asset files by deleting them, but do you mean you just deleted the whole projectsettings -folder and that fixed your strange behaviours?
I did, but I don’t recommend deleting the entire projectsettings folder as that can be a lot of havok to wreak depending on the project. That fixed the physics, but just deleting the DynamicManager.asset does that too.
I spoke too soon on the oddness with my input using the “new” Input system (EDIT: confirmed that was my own unrelated mistake). For me tho, this was at least Physics (3D).
I just came across this. Deleting the DynamicManager.asset worked. Checked version control, and nothing fundamentally changed (besides 1 name). Otherwise, it just added values (that I would think should’ve been added in the update).
Please create your own thread for your own problem.
Threads like this just accumulate different issues. The OP wasn’t discussing your reported problem here, it was related to the simulation mode not being set when updating which was fixed. The following post was completely unrelated too and was an issue that was fixed in 2D physics.
Thanks.
MelvMay - I moved the offending post, honestly I thought it was a related issue or I wouldn’t’ve posted here.
Note that if you’re saying that this thread has a clear resolution / fix to the issues raised by OP then I certainly didn’t get that from reading it, sorry.
On other forums where bugs are posted there’s usually a way to mark the bug thread resolved and add a post saying so in unambiguous terms, this might help with people accidentally necroing threads like this…
Well the thread has two completely different problems posted so this isn’t a single fix but multiple ones; hence the problem with multiple problems posts.
Agreed and this should be used as you can here too by editing the thread and marking with the RESOLVED tag as you would mark it a BUG etc but forum etiquette isn’t for everyone.
In the end, trying to stop these kinds of threads feels kind of pointless TBH as there are just so many each and every day but we try…
The main guidance here is, try to create your own thread for your own problem unless its absolutely obvious it’s the same issue and you think the thread discussion will help you solve it.
Thank you for taking the time to create your own thread!
https://discussions.unity.com/t/928269
I had a similar problem when I duplicated a Unity 2022.3.19f1 project. Some of the collision detection did not work. I solved it by changing the Project Settings-> Physics->Simulation Mode which was set to Fixed Update to something else and then back to Fixed Update.
Sorry to Necro, but encountered this in Unity 6 and couldn’t figure it out. Before deleting anything, I hopped into the Physics section of Project Settings and there’s only one field showing: “GameObject SDK”, which was set to None. I changed to PhysX and restarted and it’s working.
No idea how/why it changed, but it definitely happened following a(n unrelated?) computer freeze.