Unity 5.3.1f1 - Particle system errors - Invalid AABB result / IsFinite(d)

Thanks for the update! I'll go back to 5.3.0 then

 

Ok. fine. Fixed at least. Waiting for p1

 

Yep I just rolled back this morning, I simply cannot wait another day on development.

 

We are having this problem as well, I invested good money in the asset store and now all my particles are getting this error. This error is making the spawning system for the particles throw more errors because they can't preload properly.

 

@clearrose There are a couple of options for you listed here in the thread.

 

@MatthewW 's bandaid worked for me. My particles are spawned from a script, so I can have it auto-add that component upon creation. When the bug's fixed I'll just take it out.

 

Hi,
i am start watching the "Space Shooter" tutorial so far.
After downloading the asset(5.2) and follow the those steps of tutorial.

Finally i meet this issue=(

In this step(the link below)
http://unity3d.com/learn/tutorials/projects/space-shooter/explosions?playlist=17147

Seems passing the parameter of Quaternion to Instantiate would throw errors AABB.isdefine()

Instantiate(explosion, transform.position, transform.rotation); //Error
Instantiate(explosion, transform.position, Quaternion.identity);//No more AABB.isdefine()

 

Agh, got this too. Unplayable in the editor. Glad there is a fix

Can you go from 5.3.1 back to 5.3 without breaking anything? Or is that considered a downgrade

 

Thank you, can't wait!

 

Hi,

I update to 5.3.1 because it fixed 5.3 performance regression in dynamic batching. (749690). Does version 5.2.4 have this issue? I want go back to 5.2.4

 

May be I have related issue.
#756194

 

We have a large project that has been in the makes for a couple of years now.
Since we updated to Unity 5.3.1f1, we started getting the same errors in several scenes.
The one that caught our attention was "Invalid AABB", but it's always accompanied by things like 'IsFinite(d)'.

We were able to track the source of the issue down to GameObjects that make use of the Particle System.

I created a new project with a simple scene.
In this scene I've placed 4 copies of the same gameobject and gave them distinctive colors to recognize them more easily. The Particle Systems in these object have the same values assigned as the one causing problems in our game.


The first GameObject is positioned at 0,0,0, has no rotation and has a Scale of 0,0,0.
This one doesn't give any problems.


The second one has been moved along the z-axes and also doesn't give any problems.


The third one has been rotated, this one causes errors.


The fourth ons has been scaled up and doesn't cause any errors.

So the problem seems to lie in rotating a GameObject that has a Particle System attached to it.
When I create a new GameObject with a default Particle System and add rotation to the object, I don't get the same issues though.
When changing the simulation space to 'world', I do get the errors.

This is a snippet of the error log :


You can find the project here :
http://www.simoenss.be/Unity/ParticleErrorTest.zip

This issue is related to http://forum.unity3d.com/threads/un...errors-invalid-aabb-result-isfinite-d.374926/

 

I experience similar problems and created a bug report : http://forum.unity3d.com/threads/several-problems-with-particle-system-unity-5-3-1f1.375573/

 

@Simoense I can honestly say that i love your avatar.

 

Creating bug reports is done using the interal bug reporter tool in Unity, in the Help->Report a Bug menu. After that, post the bug number here so Unity employees can track it / escalate it further.

https://unity3d.com/unity/qa/bug-reporting

 

Thanks for reporting this, I'm merging it with the original thread to prevent duplicate topics. As far as I know this issue is now fixed and will be in 5.3.1p1

 

PLEASE WE NEED Patch 1 before the holidays!

 

I'm really sorry but it's super unlikely.

Correction below but I'm keeping the puppy.

 

I retract that, apparently, tomorrow we hope.

 

Our lord and savior of all things 2d and 3d

 

I am getting the same errors. I hope it is fixed in 5.3.1p1

 

well I found a very case-sensitive workaround until tomorrow

 

Between @Aurore and myself, you have UK and US time zones covered for the latest updates on bug fixes and patch releases over the holidays <3 I'll post here and to the 5.3 release thread when there is news. Watch for my new signature READ ME

 

Bump~~ same error

 

Great! Thanks for the follow up;

 

Had this issue from 5.3.1 too. Really scared me because it was really hard pinning the error down: the error messages are not really helpful in finding where the error occurs. Glad a fix is on the way!

 

Same, hope the patch can be rushed through for release. In meantime we don't have much of an option but to go back to 5.3

 

Hi,

I have the same issue with my game upgraded from Unity 5.3.0 to 5.3.1 f1 but I have more errors with my particle :

Invalid AABB a
IsFinite(outDistanceForSort)
IsFinite(outDistanceAlongView)

best regards,
Jeremie.

 

I just did it yesterday for a project I'm working on, nothing broke. I kept a zipped up backup in reserve just in case.

 

Wohoo! Downloading as "we speak".
Will write when I've tested it.

 

Don't know why, but my game just hangs with the 5.3.1p1 when trying to play a scene. No error messages or anything.

[Edit]
Seems to work now after restarting the editor.

The bug seems to be gone too! Great job Unity devs. <3 And merry christmas.

 

The bug with aabb messages was fixed for me too. But I have another issue, maybe worth checking.

When using particles with simulation space set to local, I "loose" particles/don't see them anymore, after few start(), stop() operations (I use self-made particle pool class, to reuse same particle effects, without having to use instantiate. The gameobj with particle systems are never destroyed, just re-positioned/rotated and ordered to play()).

If I set simulation space to world, everything works. Also, everything (local or world) worked in previous versions (versions less than 5.3).

Current workaround for me is:
1) to turn on Sub Emitters (even if I don't set any parameters, everything is set to none). Then the particle system works with simulation space set to local.
2) to use world space simulation

In both cases, the info icon shows that auto culling is disabled. So, perhaps there is a bug with aabb culling, when different simulation space is used... and particle systems are not fully reseted.

 

I was facing the same issue with 5.3.1f1 and I can confirm 5.3.1p1 works wonders! Thank you for the quick response.

 

The stack points to Render() on a reflection camera for an ocean system I didn't write (but was working fine in 5.3), I'll see what I can do

 

I have exactly the same problem.

 

@Aras Indeed, i also got that with 5.3.1 (even before patch 1) but i thought it was the same issue as with the AABB thing.
I need to set some particle systems on world space to have the play more than once when they are fired through scripting.

 

Problem fixed with Unity 5.3.1p1 patch. Thanks.

 

Thanx

 

fixed invalid aabb for me, but now I am crashing on iOS build. Doesn't happen on editor. 5.3.0 is working.

from xcode debug:

development(372,0x19f635000) malloc: *** error for object 0x10ead1020: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug

and points to line:

inline void il2cpp_codegen_marshal_free(void* ptr)
{
il2cpp::vm:latformInvoke::MarshalFree(ptr);
}

 

Albatross, this has something to do with the old error "Screen position out of view frustum" when creating the render texture (derivative from the water4 plane reflection script). I am investigating this myself but with no results till now.

Searching for it I have found that it persists along many unity versions...

 

Got it! I recreated it with water4 and a particle system

Case 756786