5.3 released

I have to ask... The new module based installation system for Unity breaks up the install so you only have to install what is necessary for the platforms you use. Sounds like a great idea.... But when installing all platforms, the installation size has jumped from 5 GB previously to 21 GB now! Not only does it take forever to download, it takes about 5x as long to install. How is this an improvement again?

 

It's double in size because of Visual Studio, which is a whopping 10+ gb, and is checked by default. Annoying. Any chance this can be unchecked by default?

Also, Ambient GI from Skybox is not fixed in 5.3.1, perhaps there wasn't enough time to hunt down this issue, but it's a serious bug. It completely breaks the lighting. See the issue here: http://forum.unity3d.com/threads/5-3-released.372567/page-2#post-2415245

 

Were any of the issues in this thread fixed in 5.3.1? I didn't see anything in the patch notes.

 

Right click C:\Program Files\Unity5.3.1f1, properties:
Size: 20.9 GB (22,510,863,384 bytes)

This does not include Visual Studio. It does include tvOS and IL2CPP support for Windows Store. Those understandably make it bigger, but > 4x the size of Unity 5.0 and > 3x the size of 5.2.2f1?

I installed all the platforms, but left out Documentation and Example Project.

19.8GB is all in Editor\Data\PlaybackEngines.

Android: 275MB
AppleTV: 1.42GB
iOS: 4.0GB
Linux Standalone: 366MB
Mac Standalone: 291MB
Metro: 12.3GB
STV: 124MB
Tizen: 36.6MB
WebGL: 349MB
Windows Standalone: 678MB

So it's mostly in Metro and iOS. More than half the install space is just Metro. Guess it has nothing to do with breaking it up into modules. Makes sense.

 

Same thing with the MonoDevelop. I manually delete it from my SSD after each installation of Unity through the Download Assistant.

 

Thanks!

 

Is my bug a known issue? Or need a bug report?

http://forum.unity3d.com/threads/5-3-released.372567/page-5#post-2429780

 

Seems like particle systems using a mesh particles are broken. Throwing lots of errors when you play the game



Edit:

I take that back. I always see these errors now, regardless of what type of renderer I have on my particle system (Mesh/Billboard/Etc). Turning off the particle system stops the errors.

 

I'm getting the same problem, turning of the particle system? what do you mean by this?

 

It seems that when I disable the game object with a particle system attached, the problem goes away.

Seems there is a surprise regression in every update. Love being kept on my toes. No I don't

 

I see, but this is not a solution as you will need the game object..Lets see what Unity say..

 

I tried to reproduce it without success, I believe that it is a memory related problem, hence it appears on my project because it's big... If there is no other solution I'll send you the entire project.
PS: great name, superpig! Matches with my t-shirt

 

Figured it out.. I am controlling the emission rate via script. I set the emission depending on the players speed. When the emission rate is set to 0 on the particle system, I get lots of errors. As soon as I start emitting particles, its fine.

Edit;

Also, simulation space must to set to 'World'.

 

Bug report with repo attached: https://fogbugz.unity3d.com/default.asp?755662_llec10kossvj9658

 

755686 UI bug: input gets masked despite canvas override ignoring it, good luck using dropdown in scrollviews, can be worked around with an ugly hack

 

is it possible to customize Visual Studio installation?

It now installs to C: drive without asking anything (cannot do that with small SSD)
and it also includes all kind of probably-not-needed tools (looks like silverlight is added, some sql server objects..)

*looks like completely uninstalling vs2013 is quite difficult..
suggestions range from complete windows reinstall to revo uninstaller..

 

Fine, one more error but not a big deal (Data not available or something)

However, I'm desperately waiting for a fix for this:


For now I won't open up Unity for a few days, thanks guys.

 

^^ thats not a bug, its a warning.

 

I know the awesome workaround for that... If you want it, PM me, and I will explain.
To everyone else: same, if you wish for a solution, PM me.

 

Or just tell us here seeing as youve already posted about it

 

I seem to be unable to get my Unity installer to download the components. It either freezes somewhere in the middle of downloading the editor, or upon restart and seemingly for the next 10 minutes, the component list itself won't even load (request timed out type errors).

Additionally, this is weird because the url's it is pointing to for component installation seem to be valid when I visit them in Safari.

Could OSX, my modem, or even my ISP itself be detecting a high volume download and then subsequently throttle my ability to download from there? Or perhaps a permissions issue on OS X? Doesn't explain why if I wait 10 minutes I can then access the component menu again, only to fail 1/2 way through the actual download.

 

Why don't you try to download from a clean virtual machine? This would help Unity's people to understand if there is a problem in the installer or if the problem is just localized in your computer.

 

I got it to work by switching over to my 2.4G network. I'm not sure what was with the 5G, as the internet seemed to work just fine, but clearly it's a local issue so you guys can ignore.

 

unity5.3 performace is 30% slower than 5.2, tested on android
anyone found this?

 

Performance in my games with unity 5.3 much better than in unity 5.2 (Android builds with standard shaders)

 

I think it is a problem. Tried it last night. Tried it right now and still waiting:

I had to re-install Unity because something was wrong with the layout colors, which never occured before.

 

Yes its first issue people noticed, please try latest 5.3.1 and see if it helps..

 

Everyhting was working fine until I got latest build tonight.

now I have errors:

Invalid AABB result
aabb.IsFinite()
Invalid AABB a x 100000 times
IsFinite(outDistanceForSort) 10000 times
IsFinite(outDistanceAlongView)

IsFinite(outDistanceForSort)
UnityEditor.DockArea:OnGUI()

IsFinite(outDistanceAlongView)
UnityEditor.DockArea:OnGUI()

And then I'm getting this nice little doosey from EASY SAVE 2: ( not 100% sure if its the plug-in that's no suddenly a problem, or Unity)
CryptographicException: Invalid input block size.
Mono.Security.Cryptography.SymmetricTransform.FinalDecrypt (System.Byte[] inputBuffer, Int32 inputOffset, Int32 inputCount) (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/Mono.Security.Cryptography/SymmetricTransform.cs:462)
Mono.Security.Cryptography.SymmetricTransform.TransformFinalBlock (System.Byte[] inputBuffer, Int32 inputOffset, Int32 inputCount) (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/Mono.Security.Cryptography/SymmetricTransform.cs:554)
System.Security.Cryptography.RijndaelManagedTransform.TransformFinalBlock (System.Byte[] inputBuffer, Int32 inputOffset, Int32 inputCount) (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/System.Security.Cryptography/RijndaelManagedTransform.cs:94)
System.Security.Cryptography.CryptoStream.FlushFinalBlock () (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/System.Security.Cryptography/CryptoStream.cs:341)
System.Security.Cryptography.CryptoStream.Close () (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/System.Security.Cryptography/CryptoStream.cs:123)
System.IO.Stream.Dispose () (at /Users/builduser/buildslave/mono-runtime-and-classlibs/build/mcs/class/corlib/System.IO/Stream.cs:98)
(wrapper remoting-invoke-with-check) System.IO.Streamispose ()
MoodkieSecurity.AESEncryptor.iDecrypt (System.Byte[] data, System.Byte[] key, System.Byte[] iv)
MoodkieSecurity.AESEncryptor.Decrypt (System.Byte[] data)
ES2Reader.GetDecryptedBytes ()
ES2Reader.GetEncryptedReader ()
ES2Reader.ReadEncrypted[Int32] (.ES2Type type)
ES2Reader.Read[Int32] (.ES2Type es2type)
ES2Reader.Read[Int32] (System.String tag)
ES2.Load[Int32] (System.String identifier)
StartLevel.Awake () (at Assets/_Scripts/StartLevel.cs:115)


NOT HAPPY JAN!

 

Hi Rich!
Robert here we were working together in Gusto and Codies

Yes, I didn't noticed any bigger issues related to this spammed error over last few months, we have noticed it first time in the 5.1.1 if I'm not mistaken. I have not seen this in logs from player outside editor, so it shouldn't do any harm or even slowdowns because of log, but it does some confusion for a moment every now and then when you play in editor and expect some issue you are trying to track down. You then have to check array of errors if the one you are looking for does not hide between those AABB. So tu summarize, its only as far harmful as obscuring editor log is ;D

Our game Thea was prized by overall stability during Early Access and we had this in our log all along. Among hundred of thousand of hours of gameplay done by our players we should have fair data base to say it might be safe to simply stop logging this error

If others can confirm the same it might help you make decision if you will be worried about "if there is any unforeseen side effect of it"

Cheers!

 

That's it! I'm tired of restarting my computer more than twice a day.
Let me uninstall it right now and get back Unity 5.2.


Let me know when it's fixed (including lightmapping).

 

Yeah I went back as well, was tempted when 5.3.1 came out, but then the complaints started right away. When 5.4 comes out, I'll try the latest 5.3

 

I'm getting a UNET crash only on iPhone that I'm not seeing with the same build on OSX. It occurs when I'm trying to join a game through matchmaking. Used to work fine on the 5.1.4 Apple TV beta.

The line is

Code (CSharp):

1. Application_set_runInBackground_m_1961756033_0(NULL/*static, unused*/, (bool)1, /*hidden argument*/NULL);

on line 34536 of Bulk_UnityEngine.Networking_0.cpp, in the "
NetworkManager_StartClient" function.

Seems to be the "SetPlayerRunInBackground" function. Not sure how to debug further, but let me know how I can provide more information?

 

So I found an easy workaround to the above crash, but it's not a permanent fix. More info in my response on this thread:

http://forum.unity3d.com/threads/unity-5-3-1-networking-unet-completely-broken-on-ios.375037/

E

 

Hello

dont know what happened wrong just update my unity version from 5.0.1 to 5.3.0

and the script started behaving very strangely...

Code (CSharp):

1.   void OnCollisionEnter2D(Collision2D other)
2.     {
3.             Destroy(gameObject);
4.             Instantiate(GM.instance.shockwave, transform.position, Quaternion.identity);
5.             GM.instance.DestroyBricks();
6.    
7.     }

in Version 5.0.1
when ball hits tiles it destroys tiles and bounce back in respective direction but after upgrading to version 5.3.0 ball destroys tiles but not bounce back , ball continue going in its respective direction....

is there is any change in OnCollisionEnter function?

 

If you had read some previous posts here...

Anyway, the reason that's behaving weirdly? Apparently, there's some screwy stuff going on in 5.3 when it comes to using 2D collision callbacks. For now, probably a good idea to roll back if you can. (Or wait for a patch release that fixes the bugs... )

 

Just got a reply from the QA team regarding my animation case 752398 (http://gfycat.com/NeighboringOddCottontail).
They said they fixed it and the problem won't appear in the upcoming Unity 5.4 release.
WHAT?
You guys introduce a bug in a major release and will only fix it in the next major release 3 months away? What if my release window is sooner than that? This is unbelievable and unacceptable!

 

I also think 5.3 was far from ready to be a marked as a stable build, specially for Android.

 

Yes this is the kind of stuff I been complaining about and deserves mention in the 5.3 performance thread.. Does it not make you just rage! If it has to be in 5.4 then a proper explanation is needed why not it can go into a patch release..

 

Guys, see this thread: http://forum.unity3d.com/threads/unity-5-3-performance-issues.372648/page-5#post-2434581

Sums up the 5.3 release, and comments from Unity on how they go about things in future.

Anyone who downloaded 5.3 should upgrade to 5.3.1 in the mean time

 

5.3.1 fixes some regressions for me. However:

- Cloudbuild still does not have 5.3.1
- We still have weird renderTexture + Stencil problems.

We basically pegged all of our builds to 5.2.x until we find a build that has all the recent problems fixed.

 

Hey Aram, just so you know, they finally pushed 5.3.1 to UCB.

 

My substance materials doesn't bake properly anymore, all maps are just replaced with random colors...



Anyone else have the same problem?

 

With bugs like these, who needs Mecanim...?

 

I did reply asking for workarounds before posting here. I really hope this is the case and the fix can be released in a patch.

 

Thanks man for narrowing down the bug. My guess is this one was tested on a hackintosh running OS 9 "Graphics: Added client/worker multithreaded rendering support to OS X Metal."
I reverted to 5.3.