Hilarious WebGL error on OS X

No need to bother with WebGL on a Mac for now, I guess

(Yes, I've reported it. But I'd be really surprised if UT don't already know this!)

 

Yup, runs well with win8.

On OSX, an empty scene errors :

Failed to run python
UnityEngine.Debug:LogError(Object)
NativeCompiler:RunProgram(ProcessStartInfo)
NativeCompiler:Execute(String, String)
UnityEditor.WebGL.Il2Cpp.EmscriptenCompiler:CompileDynamicLibrary(String, IEnumerable`1, IEnumerable`1, IEnumerable`1, IEnumerable`1)
UnityEditorInternal.IL2CPPBuilder:Run()
UnityEditorInternal.IL2CPPUtils:RunIl2Cpp(String, IIl2CppPlatformProvider, Action`1, RuntimeClassRegistry)
UnityEditor.BuildPlayerWindow:BuildPlayerAndRun()

Will find another machine w/o unity4 to try again....

 

10.7.x, 10.8.x. Which version of osx u got it running ?

 

Known bug on OSX 10.7 only.

 

So far only get it to work on win8.
Give it another run. Re-installed on OSX 10.8.5, create an empty scene and export to WebGL.
Reported to support this error :

Error building Player: Exception: Failed building WebGL Player: python "/Applications/Unity/Unity.app/Contents/PlaybackEngines/WebGLSupport/BuildTools/Emscripten/emcc" @"/Users/admin/Desktop/unity5/demoWebGL/emptyTest/Assets/../Temp/emcc_arguments.resp"

 

No joy on Windows 7 with my project either... Similar error to OP:

Code (generic):

1. Error building Player: Win32Exception: ApplicationName='"C:\Program Files\Unity 5.0.0b9\Editor\Data\PlaybackEngines\webglsupport/BuildTools/Emscripten_Win/python/2.7.5.3_64bit/python.exe"',
2.  
3. CommandLine='"C:\Program Files\Unity 5.0.0b9\Editor\Data\PlaybackEngines\webglsupport/BuildTools/Emscripten/emcc" -Oz -s NO_EXIT_RUNTIME=1 -o "D:/R&D/U5/Viewer/Assets/../Temp/StagingArea/Data\Native\UserAssembly.bc" @"C:\Users\<usr>\AppData\Local\Temp\tmp1cda64c9.tmp"',
4.  
5. CurrentDirectory='D:/R&D/U5/Viewer/Assets/../Temp/EmscriptenWork'
6.  

 

Please file bugs through the bugreporter so we can look at this.

 

It's fine on OS X 10.9, although builds kinda take forever, even with it set to the lowest optimization/fastest builds.

--Eric

 

Done! Thanks.

 

Do you have a bug # for me ?

 

Builds do take forever. Much longer than any other platform (even Flash).
Curious as if it takes as long on Windows?

 

Builds do take longer, but enabling assembly stripping in the playersettings, while doing a development build (here actually having full optimizations can make things faster), should give you the fastest turnaround time. We'll improve on this where we can, but most time is actually spent in Emscripten. They are aware of these issues and have prioritized it accordingly.

What you can use in the beta today is already an order of magnitude faster then earlier versions, so expect improvement both from us and the Emscripten parts.

 

And Yosemite, which is what I have

My issue number is 642980.

 

To everyone reporting back on WebGL in this thread; These are likely seperate issues, so please don't report every WebGL issue you have in this threads as it pollutes the thread. If it's not likely related to this issue, either start a new thread, or just file a bug report. (this doesn't mean that for you, orb, of course).

 

Still happening on beta 11. Case 645017.

 

Background : for my case, somehow it did not work on these environments 10.7.x, 10.8.x, upgraded 10.10. Simple angrybot and empty scene will trigger the phyton enscripten errors.

If you are dead beaten like me to get it working at all costs. Suggest getting a new HDD and install a clean copy of Yosemite, and it has been working like a charm for 5.0.9 and 5.0.11.

Note : it's not applicable for migrating projects from 4.x to 5.x.

 

I AM on Yosemite. I've only ever tested Unity 5 on Yosemite

 

Hi there,

I'm still getting this error in the 5.0b13 build:

Error building Player: Exception: Failed building WebGL Player: python

I've just upgraded to Yosemite and my scene is just a cube with a rotate around script.

You mention the "/proc" directory 2 posts up, I'm not sure I have that. Where would I find that exactly?

Thanks

 

In /proc - that's the full path

I very much doubt you have it, but use Terminal to make sure ("ls /").

 

Thanks for the responses, here is the Error from the console:

Error building Player: Exception: Failed building WebGL Player: python "/Applications/Unity5.0b13 /Unity.app/Contents/PlaybackEngines/WebGLSupport/BuildTools/Emscripten/emcc" @"/Users/jonathanforder/Documents/Unity3d/MobileGyroTest/Assets/../Temp/emcc_arguments.resp"

Hopefully that helps.

Thanks!

 

Ok, I've submitted a bug report.

Thanks

 

We're running into the same problem in both OSX and Windows - figure it's something to do with some of our scripts but have been unable to get any *useful* information out of the errors. We've a rather large project (at about 12GB) so submitting the project and a bug report is a bit redundant. It's a shame that the errors cannot supply the developer with more information

 

I'll try to get something up shortly - I've broken the project down into test cases so I can work through it and found one issue at least. One value:

const double M_PI = 3.1415 etc.​

which was causing the build process to crash out, but changing the name to m_Pi or MPI built successfully I'll keep picking through the project as the day goes and will report anything I do find least it's helpful

Ok, we got through the other issue the main of which turned out to be an enum that had a state as NAN, changing that to NOTN allowed the build to go through to compile. Long story short I'm not sure if the conversion from c# is case sensitive and can be a little niggly. There were also problems with some dll files and something in the StandardAssets, which in their own right were a few versions old, but I couldn't say which as I brutally stripped the project down to bare bones.

So that's the first part through I suppose, will keep posting anything I find and I do hope it helps someone out there!

 

I'm getting an error like this in beta 18:

Failed to run python

Program: python Args: "/Applications/Unity5/Unity.app/Contents/PlaybackEngines/WebGLSupport/BuildTools/Emscripten/emcc" -Oz -std=c++11 -Wno-unused-value -Wno-invalid-offsetof -I-I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/bdwgc/include" -I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/libil2cpp/include" -I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/il2cppOutput" -nostdinc -c @"/var/folders/pg/vqp5lh5x5_b_yngv7411zdv40000gp/T/tmp79cbb285.tmp"

WARNING root: [33m-I or -L of an absolute path "-I/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/libil2cpp/include" encountered. If this is to a local system header/library, it may cause problems (local system files make sense for compiling natively on your system, but not necessarily to JavaScript). Pass '-Wno-warn-absolute-paths' to emcc to hide this warning.[0m
In file included from /Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/il2cppOutput/Bulk_Arrays_10.cpp:3:
/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/libil2cpp/include/codegen/il2cpp-codegen.h:136:1: warning: function declared 'noreturn' should not return [-Winvalid-noreturn]
}
^
1 warning generated.
In file included from /Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/il2cppOutput/Bulk_Arrays_8.cpp:3:
/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/libil2cpp/include/codegen/il2cpp-codegen.h:136:1: warning: function declared 'noreturn' should not return [-Winvalid-noreturn]
}
^
1 warning generated.

rogram: python Args: "/Applications/Unity5/Unity.app/Contents/PlaybackEngines/WebGLSupport/BuildTools/Emscripten/emcc" -Oz -std=c++11 -Wno-unused-value -Wno-invalid-offsetof -I-I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/bdwgc/include" -I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/Libraries/libil2cpp/include" -I"/Users/jbooth/star_trek/client/Assets/../Temp/StagingArea/Data/il2cppOutput" -nostdinc -c @"/var/folders/pg/vqp5lh5x5_b_yngv7411zdv40000gp/T/tmp72afbf4c.tmp"

Failed to run python
UnityEngine.Debug:LogError(Object)
NativeCompiler:RunProgram(ProcessStartInfo) (at /Users/builduser/buildslave/unity/build/Editor/Mono/BuildPipeline/Il2Cpp/NativeCompiler.cs:57)
NativeCompiler:Execute(String, String) (at /Users/builduser/buildslave/unity/build/Editor/Mono/BuildPipeline/Il2Cpp/NativeCompiler.cs:23)
UnityEditor.WebGL.Il2Cpp.EmscriptenCompiler:CompileDynamicLibrary(String, IEnumerable`1, IEnumerable`1, IEnumerable`1, IEnumerable`1)
UnityEditorInternal.IL2CPPBuilder:Run() (at /Users/builduser/buildslave/unity/build/Editor/Mono/BuildPipeline/Il2Cpp/IL2CPPUtils.cs:277)
UnityEditorInternal.IL2CPPUtils:RunIl2Cpp(String, IIl2CppPlatformProvider, Action`1, RuntimeClassRegistry) (at /Users/builduser/buildslave/unity/build/Editor/Mono/BuildPipeline/Il2Cpp/IL2CPPUtils.cs:204)
UnityEditor.HostView:OnGUI()

 

I also got exactly the same issue as @jbooth using beta18 on my MacBook Pro. Just to confirm I don't have /proc directory and the level I'm building is a standard demo scene from UFPS.

I raised bug 660163 with my project attached.

A side question, is it common practice to raise bugs with a version control link instead of attachment? If so, is there a Unity support user I can invite to my github/bitbucket repository so that I don't need to make it public? Just that it took me several hours to zip up and send my 4GB project (we got pretty terrible Broadband speeds in Australia)

 

I had Beta 13 installed as of today, and my project compiled, although Arrow key input (Input.GetKey) was not recognised.

I then installed Beta 18, to see if the arrow key issue was fixed, and then I now have the same the Python issue, so it seems we took a step backwards

Failed to run "C:\Program Files\Unity 5.0.0b18\Editor\Data\PlaybackEngines\webglsupport/BuildTools/Emscripten_Win/python/2.7.5.3_64bit/python.exe"

I am using Windows 8 Bit, 64 bit editor.

 

Our reporting of what is going wrong isn't very clear, sorry.

There is no such thing as a "python" issue, although it looks that way because we haven't yet spent time of extracting a clearer error from the emscripten compiler invocation we do. What's happening is that some code that is generated from your C# code (or libraries that you pulled in) isn't compiling through the emscripten c++ compiler. Please file a bug with the project.

 

Sorry for the confusion.
I have logged a bug for my issue.

Case 660437

 

Issue also occurs in case 672778 on OSX. Reported with all debug output from the failed build.